Change Detection

Try our Chrome extension

Easily add the current web-page from your browser directly into your changedetection.io tool, more great features coming soon!

Changedetection.io needs your support!

You can help us by supporting changedetection.io on these platforms;

The more popular changedetection.io is, the more time we can dedicate to adding amazing features!

Many thanks :)

changedetection.io team

Bildirim hata ayıklama günlüğü

Wed Jul 22 04:04:30 2026 - SENDING - [{"title": "Changement d\u00e9tect\u00e9 - API OpenIA", "body": "https://developers.openai.com/api/docs/changelog a chang\u00e9e. \n--- \n            * Bulk API \n            * Overview \n            * Plugin submission errors \n            * Overview \n            * Custom Code Review rules for Codex \n            * Conversion-optimized campaigns \n---", "url": "jsons://n8n.benfou.fr/webhook/c8bb38ef-b9d8-45d0-837d-c239411db01b/?title=API%20OpenIA&avatar_url=https://raw.githubusercontent.com/dgtlmoon/changedetection.io/master/changedetectionio/static/images/avatar-256x256.png&format=html", "original_context": {"base_url": "https://changedetection.benfou.fr/", "change_datetime": "2026-07-22 04:04:17 CEST", "current_snapshot": " * Conversion-optimized campaigns", "diff": "@changed_PLACEMARKER_OPEN * Bulk API@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Overview@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Plugin submission errors@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Overview@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Custom Code Review rules for Codex@changed_PLACEMARKER_CLOSED\n@changed_into_PLACEMARKER_OPEN * Conversion-optimized campaigns@changed_into_PLACEMARKER_CLOSED", "diff_clean": "", "diff_added": "", "diff_added_clean": "", "diff_full": "", "diff_full_clean": "", "diff_patch": "", "diff_removed": "", "diff_removed_clean": "", "diff_changed_from": "", "diff_changed_to": "", "diff_url": "https://changedetection.benfou.fr/diff/40d1d040-eab5-44c8-a830-96ce77c471dd", "raw_diff": "@changed_PLACEMARKER_OPEN * Bulk API@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Overview@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Plugin submission errors@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Overview@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Custom Code Review rules for Codex@changed_PLACEMARKER_CLOSED\n@changed_into_PLACEMARKER_OPEN * Conversion-optimized campaigns@changed_into_PLACEMARKER_CLOSED", "markup_text_links_to_html_links": false, "restock": {}, "notification_timestamp": 1784685870.5184076, "prev_snapshot": " * Bulk API\n * Overview\n * Plugin submission errors\n * Overview\n * Custom Code Review rules for Codex", "preview_url": "https://changedetection.benfou.fr/preview/40d1d040-eab5-44c8-a830-96ce77c471dd", "screenshot": null, "timestamp_from": null, "timestamp_to": null, "triggered_text": "", "llm_summary": null, "llm_intent": null, "uuid": "40d1d040-eab5-44c8-a830-96ce77c471dd", "watch_mime_type": "text/html; charset=utf-8", "watch_tag": "IA", "watch_title": "API OpenIA", "watch_url": "https://developers.openai.com/api/docs/changelog", "watch_uuid": "40d1d040-eab5-44c8-a830-96ce77c471dd", "notification_urls": ["jsons://n8n.benfou.fr/webhook/c8bb38ef-b9d8-45d0-837d-c239411db01b/?title={{ watch_title|urlencode }}"], "notification_title": "Changement d\u00e9tect\u00e9 - {{watch_title}}", "notification_body": "{{watch_url}} a chang\u00e9e.\r\n---\r\n{{diff}}\r\n---\r\n", "notification_format": "htmlcolor", "_llm_result": null, "_llm_intent": "", "_llm_change_summary": "", "edit_url": "https://changedetection.benfou.fr/edit/40d1d040-eab5-44c8-a830-96ce77c471dd"}}]
Wed Jul 22 04:04:07 2026 - SENDING - [{"title": "Changement d\u00e9tect\u00e9 - API Grok", "body": "https://docs.x.ai/developers/models a chang\u00e9e. \n--- \n                  Grok Build 0.14.5 \n                grok-build-0.1 \n \n              Our coding model \u2014 trained specifically for agentic coding workflows. Use Grok Build for code. \n                grok-4.5 \n \n              Our flagship model for code and everything else: agentic tool calling, minimal hallucinations, configurable reasoning. \n                Context 256k tokens \n                Input $1.00 / 1M tokens \n                Output $2.00 / 1M tokens \n \n                  Grok 4.3 \n \n                grok-4.3 \n \n              For everything except code, audio, image, and video. The most intelligent and fastest model we\u2019ve built. \n \n                Context 1 million tokens \n                Input $1.25 / 1M tokens \n                Output $2.50 / 1M tokens \n                Context 500k tokens \n                Input $2.00 / 1M tokens \n                Output $6.00 / 1M tokens \n              View modelPlaygroundmodelTry comingin soonplayground \n---", "url": "jsons://n8n.benfou.fr/webhook/c8bb38ef-b9d8-45d0-837d-c239411db01b/?title=API%20Grok&avatar_url=https://raw.githubusercontent.com/dgtlmoon/changedetection.io/master/changedetectionio/static/images/avatar-256x256.png&format=html", "original_context": {"base_url": "https://changedetection.benfou.fr/", "change_datetime": "2026-07-22 04:03:43 CEST", "current_snapshot": " View modelPlaygroundmodelTry comingin soonplayground", "diff": "@changed_PLACEMARKER_OPEN Grok Build 0.14.5@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN grok-build-0.1@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Our coding model \u2014 trained specifically for agentic coding workflows. Use Grok Build for code.@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN grok-4.5@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Our flagship model for code and everything else: agentic tool calling, minimal hallucinations, configurable reasoning.@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Context 256k tokens@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Input $1.00 / 1M tokens@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Output $2.00 / 1M tokens@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Grok 4.3@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN grok-4.3@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN For everything except code, audio, image, and video. The most intelligent and fastest model we\u2019ve built.@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Context 1 million tokens@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Input $1.25 / 1M tokens@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Output $2.50 / 1M tokens@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Context 500k tokens@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Input $2.00 / 1M tokens@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Output $6.00 / 1M tokens@changed_PLACEMARKER_CLOSED\n@changed_into_PLACEMARKER_OPEN View modelPlaygroundmodelTry comingin soonplayground@changed_into_PLACEMARKER_CLOSED", "diff_clean": "", "diff_added": "", "diff_added_clean": "", "diff_full": "", "diff_full_clean": "", "diff_patch": "", "diff_removed": "", "diff_removed_clean": "", "diff_changed_from": "", "diff_changed_to": "", "diff_url": "https://changedetection.benfou.fr/diff/e251e15a-71cc-4dc0-b6e6-4b596730ee96", "raw_diff": "@changed_PLACEMARKER_OPEN Grok Build 0.14.5@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN grok-build-0.1@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Our coding model \u2014 trained specifically for agentic coding workflows. Use Grok Build for code.@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN grok-4.5@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Our flagship model for code and everything else: agentic tool calling, minimal hallucinations, configurable reasoning.@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Context 256k tokens@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Input $1.00 / 1M tokens@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Output $2.00 / 1M tokens@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Grok 4.3@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN grok-4.3@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN For everything except code, audio, image, and video. The most intelligent and fastest model we\u2019ve built.@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Context 1 million tokens@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Input $1.25 / 1M tokens@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Output $2.50 / 1M tokens@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Context 500k tokens@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Input $2.00 / 1M tokens@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Output $6.00 / 1M tokens@changed_PLACEMARKER_CLOSED\n@changed_into_PLACEMARKER_OPEN View modelPlaygroundmodelTry comingin soonplayground@changed_into_PLACEMARKER_CLOSED", "markup_text_links_to_html_links": false, "restock": {}, "notification_timestamp": 1784685846.80981, "prev_snapshot": " Grok Build 0.14.5\n grok-build-0.1\n\n Our coding model \u2014 trained specifically for agentic coding workflows. Use Grok Build for code.\n grok-4.5\n\n Our flagship model for code and everything else: agentic tool calling, minimal hallucinations, configurable reasoning.\n Context 256k tokens\n Input $1.00 / 1M tokens\n Output $2.00 / 1M tokens\n\n Grok 4.3\n\n grok-4.3\n\n For everything except code, audio, image, and video. The most intelligent and fastest model we\u2019ve built.\n\n Context 1 million tokens\n Input $1.25 / 1M tokens\n Output $2.50 / 1M tokens\n Context 500k tokens\n Input $2.00 / 1M tokens\n Output $6.00 / 1M tokens", "preview_url": "https://changedetection.benfou.fr/preview/e251e15a-71cc-4dc0-b6e6-4b596730ee96", "screenshot": null, "timestamp_from": null, "timestamp_to": null, "triggered_text": "", "llm_summary": null, "llm_intent": null, "uuid": "e251e15a-71cc-4dc0-b6e6-4b596730ee96", "watch_mime_type": "text/html; charset=utf-8", "watch_tag": "IA", "watch_title": "API Grok", "watch_url": "https://docs.x.ai/developers/models", "watch_uuid": "e251e15a-71cc-4dc0-b6e6-4b596730ee96", "notification_urls": ["jsons://n8n.benfou.fr/webhook/c8bb38ef-b9d8-45d0-837d-c239411db01b/?title={{ watch_title|urlencode }}"], "notification_title": "Changement d\u00e9tect\u00e9 - {{watch_title}}", "notification_body": "{{watch_url}} a chang\u00e9e.\r\n---\r\n{{diff}}\r\n---\r\n", "notification_format": "htmlcolor", "_llm_result": null, "_llm_intent": "", "_llm_change_summary": "", "edit_url": "https://changedetection.benfou.fr/edit/e251e15a-71cc-4dc0-b6e6-4b596730ee96"}}]
Tue Jul 21 21:19:33 2026 - SENDING - [{"title": "Changement d\u00e9tect\u00e9 - API OpenIA", "body": "https://developers.openai.com/api/docs/changelog a chang\u00e9e. \n--- \n            * API partner setup \n            * Conversion setup \n            * Bulk API \n            * Overview \n            * Plugin submission errors \n            * Overview \n            * Custom Code Review rules for Codex \n---", "url": "jsons://n8n.benfou.fr/webhook/c8bb38ef-b9d8-45d0-837d-c239411db01b/?title=API%20OpenIA&avatar_url=https://raw.githubusercontent.com/dgtlmoon/changedetection.io/master/changedetectionio/static/images/avatar-256x256.png&format=html", "original_context": {"base_url": "https://changedetection.benfou.fr/", "change_datetime": "2026-07-21 21:19:19 CEST", "current_snapshot": " * Bulk API\n * Overview\n * Plugin submission errors\n * Overview\n * Custom Code Review rules for Codex", "diff": "@changed_PLACEMARKER_OPEN * API partner setup@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Conversion setup@changed_PLACEMARKER_CLOSED\n@changed_into_PLACEMARKER_OPEN * Bulk API@changed_into_PLACEMARKER_CLOSED\n@changed_into_PLACEMARKER_OPEN * Overview@changed_into_PLACEMARKER_CLOSED\n@changed_into_PLACEMARKER_OPEN * Plugin submission errors@changed_into_PLACEMARKER_CLOSED\n@changed_into_PLACEMARKER_OPEN * Overview@changed_into_PLACEMARKER_CLOSED\n@changed_into_PLACEMARKER_OPEN * Custom Code Review rules for Codex@changed_into_PLACEMARKER_CLOSED", "diff_clean": "", "diff_added": "", "diff_added_clean": "", "diff_full": "", "diff_full_clean": "", "diff_patch": "", "diff_removed": "", "diff_removed_clean": "", "diff_changed_from": "", "diff_changed_to": "", "diff_url": "https://changedetection.benfou.fr/diff/40d1d040-eab5-44c8-a830-96ce77c471dd", "raw_diff": "@changed_PLACEMARKER_OPEN * API partner setup@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Conversion setup@changed_PLACEMARKER_CLOSED\n@changed_into_PLACEMARKER_OPEN * Bulk API@changed_into_PLACEMARKER_CLOSED\n@changed_into_PLACEMARKER_OPEN * Overview@changed_into_PLACEMARKER_CLOSED\n@changed_into_PLACEMARKER_OPEN * Plugin submission errors@changed_into_PLACEMARKER_CLOSED\n@changed_into_PLACEMARKER_OPEN * Overview@changed_into_PLACEMARKER_CLOSED\n@changed_into_PLACEMARKER_OPEN * Custom Code Review rules for Codex@changed_into_PLACEMARKER_CLOSED", "markup_text_links_to_html_links": false, "restock": {}, "notification_timestamp": 1784661572.866575, "prev_snapshot": " * API partner setup\n * Conversion setup", "preview_url": "https://changedetection.benfou.fr/preview/40d1d040-eab5-44c8-a830-96ce77c471dd", "screenshot": null, "timestamp_from": null, "timestamp_to": null, "triggered_text": "", "llm_summary": null, "llm_intent": null, "uuid": "40d1d040-eab5-44c8-a830-96ce77c471dd", "watch_mime_type": "text/html; charset=utf-8", "watch_tag": "IA", "watch_title": "API OpenIA", "watch_url": "https://developers.openai.com/api/docs/changelog", "watch_uuid": "40d1d040-eab5-44c8-a830-96ce77c471dd", "notification_urls": ["jsons://n8n.benfou.fr/webhook/c8bb38ef-b9d8-45d0-837d-c239411db01b/?title={{ watch_title|urlencode }}"], "notification_title": "Changement d\u00e9tect\u00e9 - {{watch_title}}", "notification_body": "{{watch_url}} a chang\u00e9e.\r\n---\r\n{{diff}}\r\n---\r\n", "notification_format": "htmlcolor", "_llm_result": null, "_llm_intent": "", "_llm_change_summary": "", "edit_url": "https://changedetection.benfou.fr/edit/40d1d040-eab5-44c8-a830-96ce77c471dd"}}]
Tue Jul 21 10:28:45 2026 - SENDING - [{"title": "Changement d\u00e9tect\u00e9 - API OpenIA", "body": "https://developers.openai.com/api/docs/changelog a chang\u00e9e. \n--- \n            * Projects, chats,Projects and taskschats \n            * Cloud FAQ \n                + Write vulnerability reports \n                + Propose security hardening \n            * ChatGPT Work Adminmode admin FAQ \n            * API partner setup \n            * Conversion setup \n---", "url": "jsons://n8n.benfou.fr/webhook/c8bb38ef-b9d8-45d0-837d-c239411db01b/?title=API%20OpenIA&avatar_url=https://raw.githubusercontent.com/dgtlmoon/changedetection.io/master/changedetectionio/static/images/avatar-256x256.png&format=html", "original_context": {"base_url": "https://changedetection.benfou.fr/", "change_datetime": "2026-07-21 10:28:28 CEST", "current_snapshot": " * API partner setup\n * Conversion setup", "diff": "@changed_PLACEMARKER_OPEN * Projects, chats,Projects and taskschats@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Cloud FAQ@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN + Write vulnerability reports@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN + Propose security hardening@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * ChatGPT Work Adminmode admin FAQ@changed_PLACEMARKER_CLOSED\n@changed_into_PLACEMARKER_OPEN * API partner setup@changed_into_PLACEMARKER_CLOSED\n@changed_into_PLACEMARKER_OPEN * Conversion setup@changed_into_PLACEMARKER_CLOSED", "diff_clean": "", "diff_added": "", "diff_added_clean": "", "diff_full": "", "diff_full_clean": "", "diff_patch": "", "diff_removed": "", "diff_removed_clean": "", "diff_changed_from": "", "diff_changed_to": "", "diff_url": "https://changedetection.benfou.fr/diff/40d1d040-eab5-44c8-a830-96ce77c471dd", "raw_diff": "@changed_PLACEMARKER_OPEN * Projects, chats,Projects and taskschats@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Cloud FAQ@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN + Write vulnerability reports@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN + Propose security hardening@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * ChatGPT Work Adminmode admin FAQ@changed_PLACEMARKER_CLOSED\n@changed_into_PLACEMARKER_OPEN * API partner setup@changed_into_PLACEMARKER_CLOSED\n@changed_into_PLACEMARKER_OPEN * Conversion setup@changed_into_PLACEMARKER_CLOSED", "markup_text_links_to_html_links": false, "restock": {}, "notification_timestamp": 1784622525.4762366, "prev_snapshot": " * Projects, chats,Projects and taskschats\n * Cloud FAQ\n + Write vulnerability reports\n + Propose security hardening\n * ChatGPT Work Adminmode admin FAQ", "preview_url": "https://changedetection.benfou.fr/preview/40d1d040-eab5-44c8-a830-96ce77c471dd", "screenshot": null, "timestamp_from": null, "timestamp_to": null, "triggered_text": "", "llm_summary": null, "llm_intent": null, "uuid": "40d1d040-eab5-44c8-a830-96ce77c471dd", "watch_mime_type": "text/html; charset=utf-8", "watch_tag": "IA", "watch_title": "API OpenIA", "watch_url": "https://developers.openai.com/api/docs/changelog", "watch_uuid": "40d1d040-eab5-44c8-a830-96ce77c471dd", "notification_urls": ["jsons://n8n.benfou.fr/webhook/c8bb38ef-b9d8-45d0-837d-c239411db01b/?title={{ watch_title|urlencode }}"], "notification_title": "Changement d\u00e9tect\u00e9 - {{watch_title}}", "notification_body": "{{watch_url}} a chang\u00e9e.\r\n---\r\n{{diff}}\r\n---\r\n", "notification_format": "htmlcolor", "_llm_result": null, "_llm_intent": "", "_llm_change_summary": "", "edit_url": "https://changedetection.benfou.fr/edit/40d1d040-eab5-44c8-a830-96ce77c471dd"}}]
Mon Jul 20 19:14:44 2026 - SENDING - [{"title": "Changement d\u00e9tect\u00e9 - Update WPML", "body": "https://wpml.org/download/wpml-multilingual-cms/?section=changelog a chang\u00e9e. \n--- \n  Version 4.9.5 \n \n  June 10, 2026 \n \n    Compatibility \n \n      * WordPress 7.0 \u2014 Fixed the Classic Translation Editor field layout, the post-language-change modal, and the language switcher block so they render correctly on WordPress 7.0. \n      * PHP 8.4 and 8.5 \u2014 Cleaned up deprecation notices and silent warnings so sites on the latest PHP versions stay clean in their error logs. \n      * Elementor \u2014 Strings sent to the Advanced Translation Editor no longer include inactive fields. \n \n    Moving and re-registering a site \n \n      * When you move a site to a new domain, WPML now applies the Copy with migration code option automatically, dismisses stale migration notices the moment the URL changes, and shows the migration banner as soon as the alias domain is reset. \n      * Restored the explanation shown when the alias domain is automatically reset, and re-enabled the manual alias-domain entry on the Troubleshooting page so site registration is reachable again on sites upgrading from 4.9.4. \n      * Stale cloned site flags are now cleared on upgrade so a partially-completed migration wizard no longer blocks subsequent runs. \n      * Redesigned the refund / unregister / replace-site-key notice as a clearer three-card layout, making it easier to remove WPML from a site you no longer use. \n \n    Advanced Translation Editor \n \n      * Slug changes made in the Advanced Translation Editor are now correctly saved back to the translated post. \n      * The lock-contention response is now a proper HTTP 423 status instead of HTTP 500, so retries from the translation service work correctly. \n      * Translation jobs are now sent in priority order \u2014 items at the top of your queue are translated first. \n      * Translations downloaded from the Advanced Translation Editor now appear in the language switcher immediately, without needing a cache flush. \n      * Translation sync error screens now include support instructions and direct links to the relevant help articles. \n      * The Translation Management job log was rebuilt on a new storage backend with much richer logging, making it easier to diagnose translation issues. \n \n    Translation Management dashboard \n \n      * The post edit screen is significantly faster on sites with orphan posts created in secondary languages \u2014 the per-render orphan-translations check has been replaced with a much lighter query, so editing posts in Gutenberg no longer pauses on large sites. \n      * Removed unused legacy tracking code and a duplicate JavaScript global from the front of the admin. \n \n    URLs, slugs, language switcher \n \n      * Plain-URL to permalink conversion now preserves the port number. \n      * The translated URL is generated the first time when translating with the native WordPress editor \u2014 no more translate twice to see the slug. \n      * Fixed translation of links at the bottom of toolset archive pages. \n      * An empty translation ID no longer accidentally creates a stray post during in-progress Classic Editor translations. \n      * Private posts no longer get redirected to a media post when they share a slug. \n      * Private posts now appear on the secondary-language blog page. \n      * Future-status posts now show the language switcher again. \n      * Language switcher block: fixed flag size in the preview area, and added explicit width and height attributes on flag images for better Core Web Vitals. \n \n    Menus, caches, and miscellaneous \n \n      * Sites running an object cache no longer lose translation menu items after Sync Menus \u2014 the term cache is now invalidated correctly. \n      * ACF Gallery Block: duplicating a page now duplicates all images, not just the first one. \n      * Cleared a false-positive SQL-injection warning from third-party security scanners. \n  Version 5.0.0-b.1 \n \n  July 20, 2026 \n \n    Translate Everything Automatically \n \n      * Translate Everything Automatically is the default translation method and can be enabled from the setup wizard. \n      * You can now choose a cutoff date for supported post types, only content created on or after that date will be included in Translate Everything Automatically; content created before it will not be translated. \n      * You can now see the cost of translation per post, and the total cost, before enabling Translate Everything Automatically. \n      * Translate Everything Automatically works only with the PTC engine. Existing sites are not affected and can continue using the Legacy engines (DeepL, Microsoft, Google). \n      * Once enabled, Translate Everything Automatically runs in the background, the browser no longer needs to stay open. \n      * Taxonomies and terms are now included in the Translate Everything Automatically process. \n      * Custom fields are included by default in the Translate Everything Automatically process. \n \n    Pay Once per Post & Language \n \n      * When Translate Everything Automatically (TEA) is active, the site is billed once per page per language. Subsequent edits to the same content are free until the content exceeds its previously paid-for length. \n      * This pricing applies only when Translate Everything Automatically is used with PTC (Private Translation Cloud), WPML\u2019s own AI translation engine. \n \n    Credits to Words Terminology Change \n \n      * The GUI now displays the actual word count for each language, and the number of words charged by the translation engine. \n \n    Translation Management Dashboard \n \n      * New section in the Translation Dashboard for Taxonomies, designed so automatic translation can be used the same way as for posts. \n      * The Translation Status section now displays details on failed translations, including a debug report and resolution steps. \n      * Previously, you had to send one or two pieces of content for a test translation before you could enable Translate Everything Automatically. Now you can activate it directly from the Dashboard, without that test step. \n \n    Translation Engine \n \n      * PTC (Private Translation Cloud), WPML\u2019s own AI translation engine, is now the recommended engine. The other three DeepL, Google, and Microsoft remain available but are labeled \u201cLegacy.\u201d \n      * Review mode defaults: \u201cNo Review\u201d is now the default review mode when PTC is enabled. For Legacy engines, the default is now \u201cReview Before Publish.\u201d \n \n    Country Pair Configuration in Language Editing \n \n      * New language settings with country-pair mapping. \n      * Improved support for creating custom languages. \n      * Displays Manual translation label for custom languages that are not mapped to a language supported by automatic translation \n \n    Translation Flow Improvements \n \n      * Supported languages are now automatically mapped, with a fallback to the base language code. For example, if a site has en-AU, the Advanced Translation Editor automatically falls back to general English and translates the content. \n      * Translators with supported language pairs can translate or take over any translation, regardless of assignment. Accessing an in-progress translation displays a takeover notice. \n \n    Improve Translations \n \n      * PTC (Private Translation Cloud), WPML\u2019s own AI translation engine, learns how you like your content translated. It automatically spots the corrections you keep making by hand. glossary terms, links, regional language variants, formality and turns them into one-time fixes. \n      * Accept a suggestion once, and PTC applies it across every current and future translation for that language pair, so you never have to make the same edit twice. \n      * A new Improve Translations tab under Translations surfaces suggestions, backed by a dashboard notice, an in-editor banner, and an email notification, so nothing slips by. \n      * Available with the PTC engine only. \n \n    Revamped, Distraction-Free Admin Interface \n \n      * Single entry point for settings, with a usable search filter. \n      * UI/UX updates to WPML\u2019s navigation (main and secondary levels), and improved cross-linking between sections. \n      * Translation Dashboard enhancements, including related views (Queue, Jobs, String Translation, etc.). \n      * Translation Dashboard improvements for clearer status indicators and next-step guidance. \n      * Regrouping of settings for better logical organization. \n      * Revised labels and descriptions for previously unclear settings. \n \n  Version 4.9.4 \n \n  May 19, 2026 \n \n    WPML 4.9.4 brings full compatibility with WordPress 7.0, addressing several changes to the block editor and core functionality introduced in this release. \n \n    WordPress 7.0 Compatibility \n \n      * The block editor now stays in modern iframe mode on WPML-enabled sites. WordPress 7.0 makes the fully iframe-based editor the new standard. On sites running an older version of WPML, the editor would silently revert to legacy mode whenever WPML was active. The Language Switcher and Navigation Language Switcher blocks have been updated to fully support Block API version 3, keeping the editor in modern mode as expected. \n      * Navigation Language Switcher \u201copen on click\u201d now works correctly on WordPress 7.0. WordPress 7.0 changed how the Navigation block stores this setting. WPML now reads the updated value correctly, so submenus respect the behaviour you configured. \n      * New WordPress 7.0 block content is now translatable. Three elements introduced or restructured in WordPress 7.0 were not appearing in the WPML translation editor. All three are now fully translatable: Icon block accessibility labels (aria-label), Navigation overlay close button text, and Tab block labels. \n      * Fixed block editor warnings and styling issues on WordPress 7.0. WPML\u2019s internal code for registering block assets has been updated to use current WordPress APIs. Block descriptions are now plain strings as WordPress 7.0 requires, and CSS loads correctly inside the iframe editor. All changes are backward compatible to WordPress 6.3. \n      * Fixed PHP deprecation notices in server logs. A PHP function deprecated in WordPress 7.0 has been replaced with the modern equivalent throughout WPML\u2019s codebase. \n \n    Other Fixes \n \n      * Fixed a blank white screen on the WPML Settings and Translation Dashboard pages in wp-admin. \n      * Fixed slow loading on the Menu Sync page. \n  Purchase \n \n  Version 4.9.3 \n \n  May 6, 2026 \n \n    WPML 4.9.3 brings a wide range of fixes and improvements \u2014 from site migration and string translation to performance, admin usability, and editor compatibility. \n \n    Features \n \n      * New filter hook for developers to return translated user meta values \n        WPML now intercepts calls to get_user_meta() and returns the translated value for the current language. This is useful when displaying user profile fields (e.g. job title, bio) on multilingual sites. To enable this behaviour, add the following to your theme or plugin: \n \n        add_filter( 'wpml_translate_get_user_meta', '__return_true' ); \n      * New filter hook to globally disable all WPML email notifications \n        WPML sends automatic emails from several internal components (translation job assignments, overdue job reports, batch summaries, etc.). Previously there was no official way to suppress them \u2014 developers had to rely on workarounds inspecting the call stack. This new hook provides a single, supported way to disable all WPML emails at once: \n \n        add_filter( 'wpml_send_email_notification', '__return_false' ); \n \n        Or selectively, based on email type and recipient: \n \n        add_filter( 'wpml_send_email_notification', function( $send, $email_type, $recipient ) { if ( $email_type === 'overdue_jobs_report' ) { return false; } return $send; }, 10, 3 ); \n \n        This is particularly useful on staging or development environments where you want to prevent emails from reaching real users. \n \n    Improvements \n \n    Migration and Setup \n \n      * Safer and More Reliable Migration Process \n        WPML now makes it safer and easier to move or copy a site from one domain to another. If something goes wrong during the migration, the wizard stays visible so users can retry, with clearer messages, better logging, and guidance on where to find the migration code. WPML also automatically fixes misconfigured alias domains, removes the unsafe \u201cMove\u201d option, and improves performance for sites using the Classic Translation Editor. \n      * Added a new option to reset the WPML Setup Wizard when it gets into a corrupted state. \n      * WPML now notifies the translation service when a site key is unregistered. \n \n    Performance \n \n      * Reduced redundant language filtering calls on sites with many ACF fields for better performance. \n      * Removed outdated WP Super Cache integration code that was no longer needed. \n \n    Admin and User Experience \n \n      * Unlocked WPML settings are now visually indicated in the admin interface. \n      * String Translation dashboard filters now show labels and default to Frontend strings. \n      * Clarified how Media Translation settings affect images and featured images. \n      * The notification about jobs that cannot be finished now links directly to the Translation Dashboard. \n \n    Fixes \n \n      * Some posts containing Base64-encoded data were not being sent for translation \n        In a previous release, we started blocking some posts because they contained Base64-encoded data, to prevent issues during the translation process. However, this also blocked some posts that could actually be translated, or where the Base64 content could simply be ignored. We\u2019ve updated how WPML handles Base64-encoded content \u2014 posts that can be safely translated are no longer blocked. A more precise validation approach is planned for an upcoming release. \n      * Fixed an issue with WPML blocks causing the Gutenberg editor to use an outdated rendering mode. \n      * Fixed an issue with Payments & Maintenance tab failing to load on sites with WordPress installed in a subdirectory. \n      * Fixed a PHP warning triggered when search queries contained array parameters. \n      * Fixed an issue with Full Site Editing category templates not applying to translated pages. \n      * Fixed an issue with translator assignment failing after a database restore or site migration. \n      * Fixed an issue with active subscriptions incorrectly shown as expired on the last day of the billing cycle. \n      * Fixed an issue with String Translation failing to load translation files due to domain name casing conflicts. \n      * Performance improvements when the Classic Translation Editor is selected. \n      * Fixed an issue with memory exhaustion errors on sites where user accounts had accented characters in their username. \n      * Fixed an issue with minimal custom field values being lost during the translation process. \n      * Fixed an issue with page preview slowness or timeouts on resource-limited servers when WPML is active. \n      * Fixed an issue with REST API requests using non-pretty URLs being incorrectly redirected. \n      * Fixed an issue with CSS changes in the WPML development project not being compiled correctly. \n      * Fixed a privacy policy link showing the original language page title instead of the translated one. \n      * Fixed an issue with the alias domain notice showing the same URL twice during site migrations. \n      * Fixed internal links using Cyrillic or non-Latin characters not being translated correctly. \n      * Fixed an issue with category and taxonomy terms not triggering the re-translation process correctly. \n      * Fixed an issue with browser language redirect not working correctly when a root page is set. \n      * Fixed an issue with the Classic Translation Editor causing PHP warnings and translation status loss on reload. \n      * Fixed an issue that was allowing posts to be accessed via the wrong category URL. \n      * Fixed an issue where String Translation file scanning was causing database corruption on large sites. \n      * Fixed an issue with the translation feedback button displaying incorrectly on Arabic (RTL) pages. \n      * Fixed a menu synchronization performance issue causing slowness or crashes on large multilingual sites. \n      * Fixed an issue with the Automatic Translation Payment & Maintenance tab failing to load on some sites. \n      * Fixed an issue with the development site banner failing the WCAG accessibility contrast requirements. \n      * Fixed an issue with internal links using www vs. non-www variants not being automatically translated. \n      * Fixed an issue with several WPML interface strings not being translated due to incorrect domain assignments. \n      * Fixed an issue with translated Elementor templates not displaying correctly on the frontend. \n      * Fixed an issue with Media Translation causing images to appear duplicated in the Media Library. \n      * Fixed an issue with the Save button remaining disabled when clearing translation context fields. \n      * Fixed an issue with website UUID mismatches causing translation service connectivity problems. \n      * Fixed WCAG accessibility issues with the legacy language switcher and SSO iframe. \n      * Fixed an issue with translator search causing heavy database load and CPU spikes. \n      * Fixed the word count calculation for translation jobs failing on Nginx/Kubernetes environments. \n      * Fixed a PHP 8.3 compatibility issue causing the Translation Dashboard to freeze. \n      * Fixed an issue with alias domain setting not being automatically removed when a domain mismatch is detected. \n      * Fixed a vulnerability in a WPML JavaScript file reported by SonarQube. \n      * Fixed an issue with class name causing issue in analytics code that could cause failures on Linux servers. \n  Purchase \n \n  Version 4.9.2.1 \n \n  April 21, 2026 \n \n    Fixes \n \n      * Fixed a security issue in WPML\u2019s Translation Feedback feature. Heads-up: we\u2019re retiring this feature in the next major WPML release because the usage has been very low. \n      * Fixed an issue where WooCommerce customers could experience a slow checkout and see a false payment error, even though their order had been successfully placed and their payment had gone through. \n      * Fixed an issue where users could see a false error when copying their site during the migration process, even though the operation had completed successfully in the background. \n  Purchase \n---", "url": "jsons://n8n.benfou.fr/webhook/c8bb38ef-b9d8-45d0-837d-c239411db01b/?title=Update%20WPML&avatar_url=https://raw.githubusercontent.com/dgtlmoon/changedetection.io/master/changedetectionio/static/images/avatar-256x256.png&format=html", "original_context": {"base_url": "https://changedetection.benfou.fr/", "change_datetime": "2026-07-20 19:14:23 CEST", "current_snapshot": " Version 5.0.0-b.1\n\n July 20, 2026\n\n Translate Everything Automatically\n\n * Translate Everything Automatically is the default translation method and can be enabled from the setup wizard.\n * You can now choose a cutoff date for supported post types, only content created on or after that date will be included in Translate Everything Automatically; content created before it will not be translated.\n * You can now see the cost of translation per post, and the total cost, before enabling Translate Everything Automatically.\n * Translate Everything Automatically works only with the PTC engine. Existing sites are not affected and can continue using the Legacy engines (DeepL, Microsoft, Google).\n * Once enabled, Translate Everything Automatically runs in the background, the browser no longer needs to stay open.\n * Taxonomies and terms are now included in the Translate Everything Automatically process.\n * Custom fields are included by default in the Translate Everything Automatically process.\n\n Pay Once per Post & Language\n\n * When Translate Everything Automatically (TEA) is active, the site is billed once per page per language. Subsequent edits to the same content are free until the content exceeds its previously paid-for length.\n * This pricing applies only when Translate Everything Automatically is used with PTC (Private Translation Cloud), WPML\u2019s own AI translation engine.\n\n Credits to Words Terminology Change\n\n * The GUI now displays the actual word count for each language, and the number of words charged by the translation engine.\n\n Translation Management Dashboard\n\n * New section in the Translation Dashboard for Taxonomies, designed so automatic translation can be used the same way as for posts.\n * The Translation Status section now displays details on failed translations, including a debug report and resolution steps.\n * Previously, you had to send one or two pieces of content for a test translation before you could enable Translate Everything Automatically. Now you can activate it directly from the Dashboard, without that test step.\n\n Translation Engine\n\n * PTC (Private Translation Cloud), WPML\u2019s own AI translation engine, is now the recommended engine. The other three DeepL, Google, and Microsoft remain available but are labeled \u201cLegacy.\u201d\n * Review mode defaults: \u201cNo Review\u201d is now the default review mode when PTC is enabled. For Legacy engines, the default is now \u201cReview Before Publish.\u201d\n\n Country Pair Configuration in Language Editing\n\n * New language settings with country-pair mapping.\n * Improved support for creating custom languages.\n * Displays Manual translation label for custom languages that are not mapped to a language supported by automatic translation\n\n Translation Flow Improvements\n\n * Supported languages are now automatically mapped, with a fallback to the base language code. For example, if a site has en-AU, the Advanced Translation Editor automatically falls back to general English and translates the content.\n * Translators with supported language pairs can translate or take over any translation, regardless of assignment. Accessing an in-progress translation displays a takeover notice.\n\n Improve Translations\n\n * PTC (Private Translation Cloud), WPML\u2019s own AI translation engine, learns how you like your content translated. It automatically spots the corrections you keep making by hand. glossary terms, links, regional language variants, formality and turns them into one-time fixes.\n * Accept a suggestion once, and PTC applies it across every current and future translation for that language pair, so you never have to make the same edit twice.\n * A new Improve Translations tab under Translations surfaces suggestions, backed by a dashboard notice, an in-editor banner, and an email notification, so nothing slips by.\n * Available with the PTC engine only.\n\n Revamped, Distraction-Free Admin Interface\n\n * Single entry point for settings, with a usable search filter.\n * UI/UX updates to WPML\u2019s navigation (main and secondary levels), and improved cross-linking between sections.\n * Translation Dashboard enhancements, including related views (Queue, Jobs, String Translation, etc.).\n * Translation Dashboard improvements for clearer status indicators and next-step guidance.\n * Regrouping of settings for better logical organization.\n * Revised labels and descriptions for previously unclear settings.\n Purchase\n", "diff": "@changed_PLACEMARKER_OPEN Version 4.9.5@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN June 10, 2026@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Compatibility@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * WordPress 7.0 \u2014 Fixed the Classic Translation Editor field layout, the post-language-change modal, and the language switcher block so they render correctly on WordPress 7.0.@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * PHP 8.4 and 8.5 \u2014 Cleaned up deprecation notices and silent warnings so sites on the latest PHP versions stay clean in their error logs.@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Elementor \u2014 Strings sent to the Advanced Translation Editor no longer include inactive fields.@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Moving and re-registering a site@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * When you move a site to a new domain, WPML now applies the Copy with migration code option automatically, dismisses stale migration notices the moment the URL changes, and shows the migration banner as soon as the alias domain is reset.@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Restored the explanation shown when the alias domain is automatically reset, and re-enabled the manual alias-domain entry on the Troubleshooting page so site registration is reachable again on sites upgrading from 4.9.4.@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Stale cloned site flags are now cleared on upgrade so a partially-completed migration wizard no longer blocks subsequent runs.@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Redesigned the refund / unregister / replace-site-key notice as a clearer three-card layout, making it easier to remove WPML from a site you no longer use.@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Advanced Translation Editor@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Slug changes made in the Advanced Translation Editor are now correctly saved back to the translated post.@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * The lock-contention response is now a proper HTTP 423 status instead of HTTP 500, so retries from the translation service work correctly.@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Translation jobs are now sent in priority order \u2014 items at the top of your queue are translated first.@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Translations downloaded from the Advanced Translation Editor now appear in the language switcher immediately, without needing a cache flush.@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Translation sync error screens now include support instructions and direct links to the relevant help articles.@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * The Translation Management job log was rebuilt on a new storage backend with much richer logging, making it easier to diagnose translation issues.@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Translation Management dashboard@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * The post edit screen is significantly faster on sites with orphan posts created in secondary languages \u2014 the per-render orphan-translations check has been replaced with a much lighter query, so editing posts in Gutenberg no longer pauses on large sites.@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Removed unused legacy tracking code and a duplicate JavaScript global from the front of the admin.@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN URLs, slugs, language switcher@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Plain-URL to permalink conversion now preserves the port number.@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * The translated URL is generated the first time when translating with the native WordPress editor \u2014 no more translate twice to see the slug.@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Fixed translation of links at the bottom of toolset archive pages.@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * An empty translation ID no longer accidentally creates a stray post during in-progress Classic Editor translations.@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Private posts no longer get redirected to a media post when they share a slug.@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Private posts now appear on the secondary-language blog page.@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Future-status posts now show the language switcher again.@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Language switcher block: fixed flag size in the preview area, and added explicit width and height attributes on flag images for better Core Web Vitals.@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Menus, caches, and miscellaneous@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Sites running an object cache no longer lose translation menu items after Sync Menus \u2014 the term cache is now invalidated correctly.@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * ACF Gallery Block: duplicating a page now duplicates all images, not just the first one.@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Cleared a false-positive SQL-injection warning from third-party security scanners.@changed_PLACEMARKER_CLOSED\n@changed_into_PLACEMARKER_OPEN Version 5.0.0-b.1@changed_into_PLACEMARKER_CLOSED\n@changed_into_PLACEMARKER_OPEN@changed_into_PLACEMARKER_CLOSED\n@changed_into_PLACEMARKER_OPEN July 20, 2026@changed_into_PLACEMARKER_CLOSED\n@changed_into_PLACEMARKER_OPEN@changed_into_PLACEMARKER_CLOSED\n@changed_into_PLACEMARKER_OPEN Translate Everything Automatically@changed_into_PLACEMARKER_CLOSED\n@changed_into_PLACEMARKER_OPEN@changed_into_PLACEMARKER_CLOSED\n@changed_into_PLACEMARKER_OPEN * Translate Everything Automatically is the default translation method and can be enabled from the setup wizard.@changed_into_PLACEMARKER_CLOSED\n@changed_into_PLACEMARKER_OPEN * You can now choose a cutoff date for supported post types, only content created on or after that date will be included in Translate Everything Automatically; content created before it will not be translated.@changed_into_PLACEMARKER_CLOSED\n@changed_into_PLACEMARKER_OPEN * You can now see the cost of translation per post, and the total cost, before enabling Translate Everything Automatically.@changed_into_PLACEMARKER_CLOSED\n@changed_into_PLACEMARKER_OPEN * Translate Everything Automatically works only with the PTC engine. Existing sites are not affected and can continue using the Legacy engines (DeepL, Microsoft, Google).@changed_into_PLACEMARKER_CLOSED\n@changed_into_PLACEMARKER_OPEN * Once enabled, Translate Everything Automatically runs in the background, the browser no longer needs to stay open.@changed_into_PLACEMARKER_CLOSED\n@changed_into_PLACEMARKER_OPEN * Taxonomies and terms are now included in the Translate Everything Automatically process.@changed_into_PLACEMARKER_CLOSED\n@changed_into_PLACEMARKER_OPEN * Custom fields are included by default in the Translate Everything Automatically process.@changed_into_PLACEMARKER_CLOSED\n@changed_into_PLACEMARKER_OPEN@changed_into_PLACEMARKER_CLOSED\n@changed_into_PLACEMARKER_OPEN Pay Once per Post & Language@changed_into_PLACEMARKER_CLOSED\n@changed_into_PLACEMARKER_OPEN@changed_into_PLACEMARKER_CLOSED\n@changed_into_PLACEMARKER_OPEN * When Translate Everything Automatically (TEA) is active, the site is billed once per page per language. Subsequent edits to the same content are free until the content exceeds its previously paid-for length.@changed_into_PLACEMARKER_CLOSED\n@changed_into_PLACEMARKER_OPEN * This pricing applies only when Translate Everything Automatically is used with PTC (Private Translation Cloud), WPML\u2019s own AI translation engine.@changed_into_PLACEMARKER_CLOSED\n@changed_into_PLACEMARKER_OPEN@changed_into_PLACEMARKER_CLOSED\n@changed_into_PLACEMARKER_OPEN Credits to Words Terminology Change@changed_into_PLACEMARKER_CLOSED\n@changed_into_PLACEMARKER_OPEN@changed_into_PLACEMARKER_CLOSED\n@changed_into_PLACEMARKER_OPEN * The GUI now displays the actual word count for each language, and the number of words charged by the translation engine.@changed_into_PLACEMARKER_CLOSED\n@changed_into_PLACEMARKER_OPEN@changed_into_PLACEMARKER_CLOSED\n@changed_into_PLACEMARKER_OPEN Translation Management Dashboard@changed_into_PLACEMARKER_CLOSED\n@changed_into_PLACEMARKER_OPEN@changed_into_PLACEMARKER_CLOSED\n@changed_into_PLACEMARKER_OPEN * New section in the Translation Dashboard for Taxonomies, designed so automatic translation can be used the same way as for posts.@changed_into_PLACEMARKER_CLOSED\n@changed_into_PLACEMARKER_OPEN * The Translation Status section now displays details on failed translations, including a debug report and resolution steps.@changed_into_PLACEMARKER_CLOSED\n@changed_into_PLACEMARKER_OPEN * Previously, you had to send one or two pieces of content for a test translation before you could enable Translate Everything Automatically. Now you can activate it directly from the Dashboard, without that test step.@changed_into_PLACEMARKER_CLOSED\n@changed_into_PLACEMARKER_OPEN@changed_into_PLACEMARKER_CLOSED\n@changed_into_PLACEMARKER_OPEN Translation Engine@changed_into_PLACEMARKER_CLOSED\n@changed_into_PLACEMARKER_OPEN@changed_into_PLACEMARKER_CLOSED\n@changed_into_PLACEMARKER_OPEN * PTC (Private Translation Cloud), WPML\u2019s own AI translation engine, is now the recommended engine. The other three DeepL, Google, and Microsoft remain available but are labeled \u201cLegacy.\u201d@changed_into_PLACEMARKER_CLOSED\n@changed_into_PLACEMARKER_OPEN * Review mode defaults: \u201cNo Review\u201d is now the default review mode when PTC is enabled. For Legacy engines, the default is now \u201cReview Before Publish.\u201d@changed_into_PLACEMARKER_CLOSED\n@changed_into_PLACEMARKER_OPEN@changed_into_PLACEMARKER_CLOSED\n@changed_into_PLACEMARKER_OPEN Country Pair Configuration in Language Editing@changed_into_PLACEMARKER_CLOSED\n@changed_into_PLACEMARKER_OPEN@changed_into_PLACEMARKER_CLOSED\n@changed_into_PLACEMARKER_OPEN * New language settings with country-pair mapping.@changed_into_PLACEMARKER_CLOSED\n@changed_into_PLACEMARKER_OPEN * Improved support for creating custom languages.@changed_into_PLACEMARKER_CLOSED\n@changed_into_PLACEMARKER_OPEN * Displays Manual translation label for custom languages that are not mapped to a language supported by automatic translation@changed_into_PLACEMARKER_CLOSED\n@changed_into_PLACEMARKER_OPEN@changed_into_PLACEMARKER_CLOSED\n@changed_into_PLACEMARKER_OPEN Translation Flow Improvements@changed_into_PLACEMARKER_CLOSED\n@changed_into_PLACEMARKER_OPEN@changed_into_PLACEMARKER_CLOSED\n@changed_into_PLACEMARKER_OPEN * Supported languages are now automatically mapped, with a fallback to the base language code. For example, if a site has en-AU, the Advanced Translation Editor automatically falls back to general English and translates the content.@changed_into_PLACEMARKER_CLOSED\n@changed_into_PLACEMARKER_OPEN * Translators with supported language pairs can translate or take over any translation, regardless of assignment. Accessing an in-progress translation displays a takeover notice.@changed_into_PLACEMARKER_CLOSED\n@changed_into_PLACEMARKER_OPEN@changed_into_PLACEMARKER_CLOSED\n@changed_into_PLACEMARKER_OPEN Improve Translations@changed_into_PLACEMARKER_CLOSED\n@changed_into_PLACEMARKER_OPEN@changed_into_PLACEMARKER_CLOSED\n@changed_into_PLACEMARKER_OPEN * PTC (Private Translation Cloud), WPML\u2019s own AI translation engine, learns how you like your content translated. It automatically spots the corrections you keep making by hand. glossary terms, links, regional language variants, formality and turns them into one-time fixes.@changed_into_PLACEMARKER_CLOSED\n@changed_into_PLACEMARKER_OPEN * Accept a suggestion once, and PTC applies it across every current and future translation for that language pair, so you never have to make the same edit twice.@changed_into_PLACEMARKER_CLOSED\n@changed_into_PLACEMARKER_OPEN * A new Improve Translations tab under Translations surfaces suggestions, backed by a dashboard notice, an in-editor banner, and an email notification, so nothing slips by.@changed_into_PLACEMARKER_CLOSED\n@changed_into_PLACEMARKER_OPEN * Available with the PTC engine only.@changed_into_PLACEMARKER_CLOSED\n@changed_into_PLACEMARKER_OPEN@changed_into_PLACEMARKER_CLOSED\n@changed_into_PLACEMARKER_OPEN Revamped, Distraction-Free Admin Interface@changed_into_PLACEMARKER_CLOSED\n@changed_into_PLACEMARKER_OPEN@changed_into_PLACEMARKER_CLOSED\n@changed_into_PLACEMARKER_OPEN * Single entry point for settings, with a usable search filter.@changed_into_PLACEMARKER_CLOSED\n@changed_into_PLACEMARKER_OPEN * UI/UX updates to WPML\u2019s navigation (main and secondary levels), and improved cross-linking between sections.@changed_into_PLACEMARKER_CLOSED\n@changed_into_PLACEMARKER_OPEN * Translation Dashboard enhancements, including related views (Queue, Jobs, String Translation, etc.).@changed_into_PLACEMARKER_CLOSED\n@changed_into_PLACEMARKER_OPEN * Translation Dashboard improvements for clearer status indicators and next-step guidance.@changed_into_PLACEMARKER_CLOSED\n@changed_into_PLACEMARKER_OPEN * Regrouping of settings for better logical organization.@changed_into_PLACEMARKER_CLOSED\n@changed_into_PLACEMARKER_OPEN * Revised labels and descriptions for previously unclear settings.@changed_into_PLACEMARKER_CLOSED\n@removed_PLACEMARKER_OPEN@removed_PLACEMARKER_CLOSED\n@removed_PLACEMARKER_OPEN Version 4.9.4@removed_PLACEMARKER_CLOSED\n@removed_PLACEMARKER_OPEN@removed_PLACEMARKER_CLOSED\n@removed_PLACEMARKER_OPEN May 19, 2026@removed_PLACEMARKER_CLOSED\n@removed_PLACEMARKER_OPEN@removed_PLACEMARKER_CLOSED\n@removed_PLACEMARKER_OPEN WPML 4.9.4 brings full compatibility with WordPress 7.0, addressing several changes to the block editor and core functionality introduced in this release.@removed_PLACEMARKER_CLOSED\n@removed_PLACEMARKER_OPEN@removed_PLACEMARKER_CLOSED\n@removed_PLACEMARKER_OPEN WordPress 7.0 Compatibility@removed_PLACEMARKER_CLOSED\n@removed_PLACEMARKER_OPEN@removed_PLACEMARKER_CLOSED\n@removed_PLACEMARKER_OPEN * The block editor now stays in modern iframe mode on WPML-enabled sites. WordPress 7.0 makes the fully iframe-based editor the new standard. On sites running an older version of WPML, the editor would silently revert to legacy mode whenever WPML was active. The Language Switcher and Navigation Language Switcher blocks have been updated to fully support Block API version 3, keeping the editor in modern mode as expected.@removed_PLACEMARKER_CLOSED\n@removed_PLACEMARKER_OPEN * Navigation Language Switcher \u201copen on click\u201d now works correctly on WordPress 7.0. WordPress 7.0 changed how the Navigation block stores this setting. WPML now reads the updated value correctly, so submenus respect the behaviour you configured.@removed_PLACEMARKER_CLOSED\n@removed_PLACEMARKER_OPEN * New WordPress 7.0 block content is now translatable. Three elements introduced or restructured in WordPress 7.0 were not appearing in the WPML translation editor. All three are now fully translatable: Icon block accessibility labels (aria-label), Navigation overlay close button text, and Tab block labels.@removed_PLACEMARKER_CLOSED\n@removed_PLACEMARKER_OPEN * Fixed block editor warnings and styling issues on WordPress 7.0. WPML\u2019s internal code for registering block assets has been updated to use current WordPress APIs. Block descriptions are now plain strings as WordPress 7.0 requires, and CSS loads correctly inside the iframe editor. All changes are backward compatible to WordPress 6.3.@removed_PLACEMARKER_CLOSED\n@removed_PLACEMARKER_OPEN * Fixed PHP deprecation notices in server logs. A PHP function deprecated in WordPress 7.0 has been replaced with the modern equivalent throughout WPML\u2019s codebase.@removed_PLACEMARKER_CLOSED\n@removed_PLACEMARKER_OPEN@removed_PLACEMARKER_CLOSED\n@removed_PLACEMARKER_OPEN Other Fixes@removed_PLACEMARKER_CLOSED\n@removed_PLACEMARKER_OPEN@removed_PLACEMARKER_CLOSED\n@removed_PLACEMARKER_OPEN * Fixed a blank white screen on the WPML Settings and Translation Dashboard pages in wp-admin.@removed_PLACEMARKER_CLOSED\n@removed_PLACEMARKER_OPEN * Fixed slow loading on the Menu Sync page.@removed_PLACEMARKER_CLOSED\n@removed_PLACEMARKER_OPEN Purchase@removed_PLACEMARKER_CLOSED\n@removed_PLACEMARKER_OPEN@removed_PLACEMARKER_CLOSED\n@removed_PLACEMARKER_OPEN Version 4.9.3@removed_PLACEMARKER_CLOSED\n@removed_PLACEMARKER_OPEN@removed_PLACEMARKER_CLOSED\n@removed_PLACEMARKER_OPEN May 6, 2026@removed_PLACEMARKER_CLOSED\n@removed_PLACEMARKER_OPEN@removed_PLACEMARKER_CLOSED\n@removed_PLACEMARKER_OPEN WPML 4.9.3 brings a wide range of fixes and improvements \u2014 from site migration and string translation to performance, admin usability, and editor compatibility.@removed_PLACEMARKER_CLOSED\n@removed_PLACEMARKER_OPEN@removed_PLACEMARKER_CLOSED\n@removed_PLACEMARKER_OPEN Features@removed_PLACEMARKER_CLOSED\n@removed_PLACEMARKER_OPEN@removed_PLACEMARKER_CLOSED\n@removed_PLACEMARKER_OPEN * New filter hook for developers to return translated user meta values@removed_PLACEMARKER_CLOSED\n@removed_PLACEMARKER_OPEN WPML now intercepts calls to get_user_meta() and returns the translated value for the current language. This is useful when displaying user profile fields (e.g. job title, bio) on multilingual sites. To enable this behaviour, add the following to your theme or plugin:@removed_PLACEMARKER_CLOSED\n@removed_PLACEMARKER_OPEN@removed_PLACEMARKER_CLOSED\n@removed_PLACEMARKER_OPEN add_filter( 'wpml_translate_get_user_meta', '__return_true' );@removed_PLACEMARKER_CLOSED\n@removed_PLACEMARKER_OPEN * New filter hook to globally disable all WPML email notifications@removed_PLACEMARKER_CLOSED\n@removed_PLACEMARKER_OPEN WPML sends automatic emails from several internal components (translation job assignments, overdue job reports, batch summaries, etc.). Previously there was no official way to suppress them \u2014 developers had to rely on workarounds inspecting the call stack. This new hook provides a single, supported way to disable all WPML emails at once:@removed_PLACEMARKER_CLOSED\n@removed_PLACEMARKER_OPEN@removed_PLACEMARKER_CLOSED\n@removed_PLACEMARKER_OPEN add_filter( 'wpml_send_email_notification', '__return_false' );@removed_PLACEMARKER_CLOSED\n@removed_PLACEMARKER_OPEN@removed_PLACEMARKER_CLOSED\n@removed_PLACEMARKER_OPEN Or selectively, based on email type and recipient:@removed_PLACEMARKER_CLOSED\n@removed_PLACEMARKER_OPEN@removed_PLACEMARKER_CLOSED\n@removed_PLACEMARKER_OPEN add_filter( 'wpml_send_email_notification', function( $send, $email_type, $recipient ) { if ( $email_type === 'overdue_jobs_report' ) { return false; } return $send; }, 10, 3 );@removed_PLACEMARKER_CLOSED\n@removed_PLACEMARKER_OPEN@removed_PLACEMARKER_CLOSED\n@removed_PLACEMARKER_OPEN This is particularly useful on staging or development environments where you want to prevent emails from reaching real users.@removed_PLACEMARKER_CLOSED\n@removed_PLACEMARKER_OPEN@removed_PLACEMARKER_CLOSED\n@removed_PLACEMARKER_OPEN Improvements@removed_PLACEMARKER_CLOSED\n@removed_PLACEMARKER_OPEN@removed_PLACEMARKER_CLOSED\n@removed_PLACEMARKER_OPEN Migration and Setup@removed_PLACEMARKER_CLOSED\n@removed_PLACEMARKER_OPEN@removed_PLACEMARKER_CLOSED\n@removed_PLACEMARKER_OPEN * Safer and More Reliable Migration Process@removed_PLACEMARKER_CLOSED\n@removed_PLACEMARKER_OPEN WPML now makes it safer and easier to move or copy a site from one domain to another. If something goes wrong during the migration, the wizard stays visible so users can retry, with clearer messages, better logging, and guidance on where to find the migration code. WPML also automatically fixes misconfigured alias domains, removes the unsafe \u201cMove\u201d option, and improves performance for sites using the Classic Translation Editor.@removed_PLACEMARKER_CLOSED\n@removed_PLACEMARKER_OPEN * Added a new option to reset the WPML Setup Wizard when it gets into a corrupted state.@removed_PLACEMARKER_CLOSED\n@removed_PLACEMARKER_OPEN * WPML now notifies the translation service when a site key is unregistered.@removed_PLACEMARKER_CLOSED\n@removed_PLACEMARKER_OPEN@removed_PLACEMARKER_CLOSED\n@removed_PLACEMARKER_OPEN Performance@removed_PLACEMARKER_CLOSED\n@removed_PLACEMARKER_OPEN@removed_PLACEMARKER_CLOSED\n@removed_PLACEMARKER_OPEN * Reduced redundant language filtering calls on sites with many ACF fields for better performance.@removed_PLACEMARKER_CLOSED\n@removed_PLACEMARKER_OPEN * Removed outdated WP Super Cache integration code that was no longer needed.@removed_PLACEMARKER_CLOSED\n@removed_PLACEMARKER_OPEN@removed_PLACEMARKER_CLOSED\n@removed_PLACEMARKER_OPEN Admin and User Experience@removed_PLACEMARKER_CLOSED\n@removed_PLACEMARKER_OPEN@removed_PLACEMARKER_CLOSED\n@removed_PLACEMARKER_OPEN * Unlocked WPML settings are now visually indicated in the admin interface.@removed_PLACEMARKER_CLOSED\n@removed_PLACEMARKER_OPEN * String Translation dashboard filters now show labels and default to Frontend strings.@removed_PLACEMARKER_CLOSED\n@removed_PLACEMARKER_OPEN * Clarified how Media Translation settings affect images and featured images.@removed_PLACEMARKER_CLOSED\n@removed_PLACEMARKER_OPEN * The notification about jobs that cannot be finished now links directly to the Translation Dashboard.@removed_PLACEMARKER_CLOSED\n@removed_PLACEMARKER_OPEN@removed_PLACEMARKER_CLOSED\n@removed_PLACEMARKER_OPEN Fixes@removed_PLACEMARKER_CLOSED\n@removed_PLACEMARKER_OPEN@removed_PLACEMARKER_CLOSED\n@removed_PLACEMARKER_OPEN * Some posts containing Base64-encoded data were not being sent for translation@removed_PLACEMARKER_CLOSED\n@removed_PLACEMARKER_OPEN In a previous release, we started blocking some posts because they contained Base64-encoded data, to prevent issues during the translation process. However, this also blocked some posts that could actually be translated, or where the Base64 content could simply be ignored. We\u2019ve updated how WPML handles Base64-encoded content \u2014 posts that can be safely translated are no longer blocked. A more precise validation approach is planned for an upcoming release.@removed_PLACEMARKER_CLOSED\n@removed_PLACEMARKER_OPEN * Fixed an issue with WPML blocks causing the Gutenberg editor to use an outdated rendering mode.@removed_PLACEMARKER_CLOSED\n@removed_PLACEMARKER_OPEN * Fixed an issue with Payments & Maintenance tab failing to load on sites with WordPress installed in a subdirectory.@removed_PLACEMARKER_CLOSED\n@removed_PLACEMARKER_OPEN * Fixed a PHP warning triggered when search queries contained array parameters.@removed_PLACEMARKER_CLOSED\n@removed_PLACEMARKER_OPEN * Fixed an issue with Full Site Editing category templates not applying to translated pages.@removed_PLACEMARKER_CLOSED\n@removed_PLACEMARKER_OPEN * Fixed an issue with translator assignment failing after a database restore or site migration.@removed_PLACEMARKER_CLOSED\n@removed_PLACEMARKER_OPEN * Fixed an issue with active subscriptions incorrectly shown as expired on the last day of the billing cycle.@removed_PLACEMARKER_CLOSED\n@removed_PLACEMARKER_OPEN * Fixed an issue with String Translation failing to load translation files due to domain name casing conflicts.@removed_PLACEMARKER_CLOSED\n@removed_PLACEMARKER_OPEN * Performance improvements when the Classic Translation Editor is selected.@removed_PLACEMARKER_CLOSED\n@removed_PLACEMARKER_OPEN * Fixed an issue with memory exhaustion errors on sites where user accounts had accented characters in their username.@removed_PLACEMARKER_CLOSED\n@removed_PLACEMARKER_OPEN * Fixed an issue with minimal custom field values being lost during the translation process.@removed_PLACEMARKER_CLOSED\n@removed_PLACEMARKER_OPEN * Fixed an issue with page preview slowness or timeouts on resource-limited servers when WPML is active.@removed_PLACEMARKER_CLOSED\n@removed_PLACEMARKER_OPEN * Fixed an issue with REST API requests using non-pretty URLs being incorrectly redirected.@removed_PLACEMARKER_CLOSED\n@removed_PLACEMARKER_OPEN * Fixed an issue with CSS changes in the WPML development project not being compiled correctly.@removed_PLACEMARKER_CLOSED\n@removed_PLACEMARKER_OPEN * Fixed a privacy policy link showing the original language page title instead of the translated one.@removed_PLACEMARKER_CLOSED\n@removed_PLACEMARKER_OPEN * Fixed an issue with the alias domain notice showing the same URL twice during site migrations.@removed_PLACEMARKER_CLOSED\n@removed_PLACEMARKER_OPEN * Fixed internal links using Cyrillic or non-Latin characters not being translated correctly.@removed_PLACEMARKER_CLOSED\n@removed_PLACEMARKER_OPEN * Fixed an issue with category and taxonomy terms not triggering the re-translation process correctly.@removed_PLACEMARKER_CLOSED\n@removed_PLACEMARKER_OPEN * Fixed an issue with browser language redirect not working correctly when a root page is set.@removed_PLACEMARKER_CLOSED\n@removed_PLACEMARKER_OPEN * Fixed an issue with the Classic Translation Editor causing PHP warnings and translation status loss on reload.@removed_PLACEMARKER_CLOSED\n@removed_PLACEMARKER_OPEN * Fixed an issue that was allowing posts to be accessed via the wrong category URL.@removed_PLACEMARKER_CLOSED\n@removed_PLACEMARKER_OPEN * Fixed an issue where String Translation file scanning was causing database corruption on large sites.@removed_PLACEMARKER_CLOSED\n@removed_PLACEMARKER_OPEN * Fixed an issue with the translation feedback button displaying incorrectly on Arabic (RTL) pages.@removed_PLACEMARKER_CLOSED\n@removed_PLACEMARKER_OPEN * Fixed a menu synchronization performance issue causing slowness or crashes on large multilingual sites.@removed_PLACEMARKER_CLOSED\n@removed_PLACEMARKER_OPEN * Fixed an issue with the Automatic Translation Payment & Maintenance tab failing to load on some sites.@removed_PLACEMARKER_CLOSED\n@removed_PLACEMARKER_OPEN * Fixed an issue with the development site banner failing the WCAG accessibility contrast requirements.@removed_PLACEMARKER_CLOSED\n@removed_PLACEMARKER_OPEN * Fixed an issue with internal links using www vs. non-www variants not being automatically translated.@removed_PLACEMARKER_CLOSED\n@removed_PLACEMARKER_OPEN * Fixed an issue with several WPML interface strings not being translated due to incorrect domain assignments.@removed_PLACEMARKER_CLOSED\n@removed_PLACEMARKER_OPEN * Fixed an issue with translated Elementor templates not displaying correctly on the frontend.@removed_PLACEMARKER_CLOSED\n@removed_PLACEMARKER_OPEN * Fixed an issue with Media Translation causing images to appear duplicated in the Media Library.@removed_PLACEMARKER_CLOSED\n@removed_PLACEMARKER_OPEN * Fixed an issue with the Save button remaining disabled when clearing translation context fields.@removed_PLACEMARKER_CLOSED\n@removed_PLACEMARKER_OPEN * Fixed an issue with website UUID mismatches causing translation service connectivity problems.@removed_PLACEMARKER_CLOSED\n@removed_PLACEMARKER_OPEN * Fixed WCAG accessibility issues with the legacy language switcher and SSO iframe.@removed_PLACEMARKER_CLOSED\n@removed_PLACEMARKER_OPEN * Fixed an issue with translator search causing heavy database load and CPU spikes.@removed_PLACEMARKER_CLOSED\n@removed_PLACEMARKER_OPEN * Fixed the word count calculation for translation jobs failing on Nginx/Kubernetes environments.@removed_PLACEMARKER_CLOSED\n@removed_PLACEMARKER_OPEN * Fixed a PHP 8.3 compatibility issue causing the Translation Dashboard to freeze.@removed_PLACEMARKER_CLOSED\n@removed_PLACEMARKER_OPEN * Fixed an issue with alias domain setting not being automatically removed when a domain mismatch is detected.@removed_PLACEMARKER_CLOSED\n@removed_PLACEMARKER_OPEN * Fixed a vulnerability in a WPML JavaScript file reported by SonarQube.@removed_PLACEMARKER_CLOSED\n@removed_PLACEMARKER_OPEN * Fixed an issue with class name causing issue in analytics code that could cause failures on Linux servers.@removed_PLACEMARKER_CLOSED\n@removed_PLACEMARKER_OPEN Purchase@removed_PLACEMARKER_CLOSED\n@removed_PLACEMARKER_OPEN@removed_PLACEMARKER_CLOSED\n@removed_PLACEMARKER_OPEN Version 4.9.2.1@removed_PLACEMARKER_CLOSED\n@removed_PLACEMARKER_OPEN@removed_PLACEMARKER_CLOSED\n@removed_PLACEMARKER_OPEN April 21, 2026@removed_PLACEMARKER_CLOSED\n@removed_PLACEMARKER_OPEN@removed_PLACEMARKER_CLOSED\n@removed_PLACEMARKER_OPEN Fixes@removed_PLACEMARKER_CLOSED\n@removed_PLACEMARKER_OPEN@removed_PLACEMARKER_CLOSED\n@removed_PLACEMARKER_OPEN * Fixed a security issue in WPML\u2019s Translation Feedback feature. Heads-up: we\u2019re retiring this feature in the next major WPML release because the usage has been very low.@removed_PLACEMARKER_CLOSED\n@removed_PLACEMARKER_OPEN * Fixed an issue where WooCommerce customers could experience a slow checkout and see a false payment error, even though their order had been successfully placed and their payment had gone through.@removed_PLACEMARKER_CLOSED\n@removed_PLACEMARKER_OPEN * Fixed an issue where users could see a false error when copying their site during the migration process, even though the operation had completed successfully in the background.@removed_PLACEMARKER_CLOSED\n@removed_PLACEMARKER_OPEN Purchase@removed_PLACEMARKER_CLOSED", "diff_clean": "", "diff_added": "", "diff_added_clean": "", "diff_full": "", "diff_full_clean": "", "diff_patch": "", "diff_removed": "", "diff_removed_clean": "", "diff_changed_from": "", "diff_changed_to": "", "diff_url": "https://changedetection.benfou.fr/diff/83594c1a-d307-48ee-a2dd-1586170fc4f4", "raw_diff": "@changed_PLACEMARKER_OPEN Version 4.9.5@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN June 10, 2026@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Compatibility@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * WordPress 7.0 \u2014 Fixed the Classic Translation Editor field layout, the post-language-change modal, and the language switcher block so they render correctly on WordPress 7.0.@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * PHP 8.4 and 8.5 \u2014 Cleaned up deprecation notices and silent warnings so sites on the latest PHP versions stay clean in their error logs.@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Elementor \u2014 Strings sent to the Advanced Translation Editor no longer include inactive fields.@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Moving and re-registering a site@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * When you move a site to a new domain, WPML now applies the Copy with migration code option automatically, dismisses stale migration notices the moment the URL changes, and shows the migration banner as soon as the alias domain is reset.@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Restored the explanation shown when the alias domain is automatically reset, and re-enabled the manual alias-domain entry on the Troubleshooting page so site registration is reachable again on sites upgrading from 4.9.4.@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Stale cloned site flags are now cleared on upgrade so a partially-completed migration wizard no longer blocks subsequent runs.@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Redesigned the refund / unregister / replace-site-key notice as a clearer three-card layout, making it easier to remove WPML from a site you no longer use.@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Advanced Translation Editor@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Slug changes made in the Advanced Translation Editor are now correctly saved back to the translated post.@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * The lock-contention response is now a proper HTTP 423 status instead of HTTP 500, so retries from the translation service work correctly.@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Translation jobs are now sent in priority order \u2014 items at the top of your queue are translated first.@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Translations downloaded from the Advanced Translation Editor now appear in the language switcher immediately, without needing a cache flush.@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Translation sync error screens now include support instructions and direct links to the relevant help articles.@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * The Translation Management job log was rebuilt on a new storage backend with much richer logging, making it easier to diagnose translation issues.@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Translation Management dashboard@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * The post edit screen is significantly faster on sites with orphan posts created in secondary languages \u2014 the per-render orphan-translations check has been replaced with a much lighter query, so editing posts in Gutenberg no longer pauses on large sites.@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Removed unused legacy tracking code and a duplicate JavaScript global from the front of the admin.@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN URLs, slugs, language switcher@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Plain-URL to permalink conversion now preserves the port number.@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * The translated URL is generated the first time when translating with the native WordPress editor \u2014 no more translate twice to see the slug.@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Fixed translation of links at the bottom of toolset archive pages.@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * An empty translation ID no longer accidentally creates a stray post during in-progress Classic Editor translations.@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Private posts no longer get redirected to a media post when they share a slug.@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Private posts now appear on the secondary-language blog page.@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Future-status posts now show the language switcher again.@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Language switcher block: fixed flag size in the preview area, and added explicit width and height attributes on flag images for better Core Web Vitals.@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Menus, caches, and miscellaneous@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Sites running an object cache no longer lose translation menu items after Sync Menus \u2014 the term cache is now invalidated correctly.@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * ACF Gallery Block: duplicating a page now duplicates all images, not just the first one.@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Cleared a false-positive SQL-injection warning from third-party security scanners.@changed_PLACEMARKER_CLOSED\n@changed_into_PLACEMARKER_OPEN Version 5.0.0-b.1@changed_into_PLACEMARKER_CLOSED\n@changed_into_PLACEMARKER_OPEN@changed_into_PLACEMARKER_CLOSED\n@changed_into_PLACEMARKER_OPEN July 20, 2026@changed_into_PLACEMARKER_CLOSED\n@changed_into_PLACEMARKER_OPEN@changed_into_PLACEMARKER_CLOSED\n@changed_into_PLACEMARKER_OPEN Translate Everything Automatically@changed_into_PLACEMARKER_CLOSED\n@changed_into_PLACEMARKER_OPEN@changed_into_PLACEMARKER_CLOSED\n@changed_into_PLACEMARKER_OPEN * Translate Everything Automatically is the default translation method and can be enabled from the setup wizard.@changed_into_PLACEMARKER_CLOSED\n@changed_into_PLACEMARKER_OPEN * You can now choose a cutoff date for supported post types, only content created on or after that date will be included in Translate Everything Automatically; content created before it will not be translated.@changed_into_PLACEMARKER_CLOSED\n@changed_into_PLACEMARKER_OPEN * You can now see the cost of translation per post, and the total cost, before enabling Translate Everything Automatically.@changed_into_PLACEMARKER_CLOSED\n@changed_into_PLACEMARKER_OPEN * Translate Everything Automatically works only with the PTC engine. Existing sites are not affected and can continue using the Legacy engines (DeepL, Microsoft, Google).@changed_into_PLACEMARKER_CLOSED\n@changed_into_PLACEMARKER_OPEN * Once enabled, Translate Everything Automatically runs in the background, the browser no longer needs to stay open.@changed_into_PLACEMARKER_CLOSED\n@changed_into_PLACEMARKER_OPEN * Taxonomies and terms are now included in the Translate Everything Automatically process.@changed_into_PLACEMARKER_CLOSED\n@changed_into_PLACEMARKER_OPEN * Custom fields are included by default in the Translate Everything Automatically process.@changed_into_PLACEMARKER_CLOSED\n@changed_into_PLACEMARKER_OPEN@changed_into_PLACEMARKER_CLOSED\n@changed_into_PLACEMARKER_OPEN Pay Once per Post & Language@changed_into_PLACEMARKER_CLOSED\n@changed_into_PLACEMARKER_OPEN@changed_into_PLACEMARKER_CLOSED\n@changed_into_PLACEMARKER_OPEN * When Translate Everything Automatically (TEA) is active, the site is billed once per page per language. Subsequent edits to the same content are free until the content exceeds its previously paid-for length.@changed_into_PLACEMARKER_CLOSED\n@changed_into_PLACEMARKER_OPEN * This pricing applies only when Translate Everything Automatically is used with PTC (Private Translation Cloud), WPML\u2019s own AI translation engine.@changed_into_PLACEMARKER_CLOSED\n@changed_into_PLACEMARKER_OPEN@changed_into_PLACEMARKER_CLOSED\n@changed_into_PLACEMARKER_OPEN Credits to Words Terminology Change@changed_into_PLACEMARKER_CLOSED\n@changed_into_PLACEMARKER_OPEN@changed_into_PLACEMARKER_CLOSED\n@changed_into_PLACEMARKER_OPEN * The GUI now displays the actual word count for each language, and the number of words charged by the translation engine.@changed_into_PLACEMARKER_CLOSED\n@changed_into_PLACEMARKER_OPEN@changed_into_PLACEMARKER_CLOSED\n@changed_into_PLACEMARKER_OPEN Translation Management Dashboard@changed_into_PLACEMARKER_CLOSED\n@changed_into_PLACEMARKER_OPEN@changed_into_PLACEMARKER_CLOSED\n@changed_into_PLACEMARKER_OPEN * New section in the Translation Dashboard for Taxonomies, designed so automatic translation can be used the same way as for posts.@changed_into_PLACEMARKER_CLOSED\n@changed_into_PLACEMARKER_OPEN * The Translation Status section now displays details on failed translations, including a debug report and resolution steps.@changed_into_PLACEMARKER_CLOSED\n@changed_into_PLACEMARKER_OPEN * Previously, you had to send one or two pieces of content for a test translation before you could enable Translate Everything Automatically. Now you can activate it directly from the Dashboard, without that test step.@changed_into_PLACEMARKER_CLOSED\n@changed_into_PLACEMARKER_OPEN@changed_into_PLACEMARKER_CLOSED\n@changed_into_PLACEMARKER_OPEN Translation Engine@changed_into_PLACEMARKER_CLOSED\n@changed_into_PLACEMARKER_OPEN@changed_into_PLACEMARKER_CLOSED\n@changed_into_PLACEMARKER_OPEN * PTC (Private Translation Cloud), WPML\u2019s own AI translation engine, is now the recommended engine. The other three DeepL, Google, and Microsoft remain available but are labeled \u201cLegacy.\u201d@changed_into_PLACEMARKER_CLOSED\n@changed_into_PLACEMARKER_OPEN * Review mode defaults: \u201cNo Review\u201d is now the default review mode when PTC is enabled. For Legacy engines, the default is now \u201cReview Before Publish.\u201d@changed_into_PLACEMARKER_CLOSED\n@changed_into_PLACEMARKER_OPEN@changed_into_PLACEMARKER_CLOSED\n@changed_into_PLACEMARKER_OPEN Country Pair Configuration in Language Editing@changed_into_PLACEMARKER_CLOSED\n@changed_into_PLACEMARKER_OPEN@changed_into_PLACEMARKER_CLOSED\n@changed_into_PLACEMARKER_OPEN * New language settings with country-pair mapping.@changed_into_PLACEMARKER_CLOSED\n@changed_into_PLACEMARKER_OPEN * Improved support for creating custom languages.@changed_into_PLACEMARKER_CLOSED\n@changed_into_PLACEMARKER_OPEN * Displays Manual translation label for custom languages that are not mapped to a language supported by automatic translation@changed_into_PLACEMARKER_CLOSED\n@changed_into_PLACEMARKER_OPEN@changed_into_PLACEMARKER_CLOSED\n@changed_into_PLACEMARKER_OPEN Translation Flow Improvements@changed_into_PLACEMARKER_CLOSED\n@changed_into_PLACEMARKER_OPEN@changed_into_PLACEMARKER_CLOSED\n@changed_into_PLACEMARKER_OPEN * Supported languages are now automatically mapped, with a fallback to the base language code. For example, if a site has en-AU, the Advanced Translation Editor automatically falls back to general English and translates the content.@changed_into_PLACEMARKER_CLOSED\n@changed_into_PLACEMARKER_OPEN * Translators with supported language pairs can translate or take over any translation, regardless of assignment. Accessing an in-progress translation displays a takeover notice.@changed_into_PLACEMARKER_CLOSED\n@changed_into_PLACEMARKER_OPEN@changed_into_PLACEMARKER_CLOSED\n@changed_into_PLACEMARKER_OPEN Improve Translations@changed_into_PLACEMARKER_CLOSED\n@changed_into_PLACEMARKER_OPEN@changed_into_PLACEMARKER_CLOSED\n@changed_into_PLACEMARKER_OPEN * PTC (Private Translation Cloud), WPML\u2019s own AI translation engine, learns how you like your content translated. It automatically spots the corrections you keep making by hand. glossary terms, links, regional language variants, formality and turns them into one-time fixes.@changed_into_PLACEMARKER_CLOSED\n@changed_into_PLACEMARKER_OPEN * Accept a suggestion once, and PTC applies it across every current and future translation for that language pair, so you never have to make the same edit twice.@changed_into_PLACEMARKER_CLOSED\n@changed_into_PLACEMARKER_OPEN * A new Improve Translations tab under Translations surfaces suggestions, backed by a dashboard notice, an in-editor banner, and an email notification, so nothing slips by.@changed_into_PLACEMARKER_CLOSED\n@changed_into_PLACEMARKER_OPEN * Available with the PTC engine only.@changed_into_PLACEMARKER_CLOSED\n@changed_into_PLACEMARKER_OPEN@changed_into_PLACEMARKER_CLOSED\n@changed_into_PLACEMARKER_OPEN Revamped, Distraction-Free Admin Interface@changed_into_PLACEMARKER_CLOSED\n@changed_into_PLACEMARKER_OPEN@changed_into_PLACEMARKER_CLOSED\n@changed_into_PLACEMARKER_OPEN * Single entry point for settings, with a usable search filter.@changed_into_PLACEMARKER_CLOSED\n@changed_into_PLACEMARKER_OPEN * UI/UX updates to WPML\u2019s navigation (main and secondary levels), and improved cross-linking between sections.@changed_into_PLACEMARKER_CLOSED\n@changed_into_PLACEMARKER_OPEN * Translation Dashboard enhancements, including related views (Queue, Jobs, String Translation, etc.).@changed_into_PLACEMARKER_CLOSED\n@changed_into_PLACEMARKER_OPEN * Translation Dashboard improvements for clearer status indicators and next-step guidance.@changed_into_PLACEMARKER_CLOSED\n@changed_into_PLACEMARKER_OPEN * Regrouping of settings for better logical organization.@changed_into_PLACEMARKER_CLOSED\n@changed_into_PLACEMARKER_OPEN * Revised labels and descriptions for previously unclear settings.@changed_into_PLACEMARKER_CLOSED\n@removed_PLACEMARKER_OPEN@removed_PLACEMARKER_CLOSED\n@removed_PLACEMARKER_OPEN Version 4.9.4@removed_PLACEMARKER_CLOSED\n@removed_PLACEMARKER_OPEN@removed_PLACEMARKER_CLOSED\n@removed_PLACEMARKER_OPEN May 19, 2026@removed_PLACEMARKER_CLOSED\n@removed_PLACEMARKER_OPEN@removed_PLACEMARKER_CLOSED\n@removed_PLACEMARKER_OPEN WPML 4.9.4 brings full compatibility with WordPress 7.0, addressing several changes to the block editor and core functionality introduced in this release.@removed_PLACEMARKER_CLOSED\n@removed_PLACEMARKER_OPEN@removed_PLACEMARKER_CLOSED\n@removed_PLACEMARKER_OPEN WordPress 7.0 Compatibility@removed_PLACEMARKER_CLOSED\n@removed_PLACEMARKER_OPEN@removed_PLACEMARKER_CLOSED\n@removed_PLACEMARKER_OPEN * The block editor now stays in modern iframe mode on WPML-enabled sites. WordPress 7.0 makes the fully iframe-based editor the new standard. On sites running an older version of WPML, the editor would silently revert to legacy mode whenever WPML was active. The Language Switcher and Navigation Language Switcher blocks have been updated to fully support Block API version 3, keeping the editor in modern mode as expected.@removed_PLACEMARKER_CLOSED\n@removed_PLACEMARKER_OPEN * Navigation Language Switcher \u201copen on click\u201d now works correctly on WordPress 7.0. WordPress 7.0 changed how the Navigation block stores this setting. WPML now reads the updated value correctly, so submenus respect the behaviour you configured.@removed_PLACEMARKER_CLOSED\n@removed_PLACEMARKER_OPEN * New WordPress 7.0 block content is now translatable. Three elements introduced or restructured in WordPress 7.0 were not appearing in the WPML translation editor. All three are now fully translatable: Icon block accessibility labels (aria-label), Navigation overlay close button text, and Tab block labels.@removed_PLACEMARKER_CLOSED\n@removed_PLACEMARKER_OPEN * Fixed block editor warnings and styling issues on WordPress 7.0. WPML\u2019s internal code for registering block assets has been updated to use current WordPress APIs. Block descriptions are now plain strings as WordPress 7.0 requires, and CSS loads correctly inside the iframe editor. All changes are backward compatible to WordPress 6.3.@removed_PLACEMARKER_CLOSED\n@removed_PLACEMARKER_OPEN * Fixed PHP deprecation notices in server logs. A PHP function deprecated in WordPress 7.0 has been replaced with the modern equivalent throughout WPML\u2019s codebase.@removed_PLACEMARKER_CLOSED\n@removed_PLACEMARKER_OPEN@removed_PLACEMARKER_CLOSED\n@removed_PLACEMARKER_OPEN Other Fixes@removed_PLACEMARKER_CLOSED\n@removed_PLACEMARKER_OPEN@removed_PLACEMARKER_CLOSED\n@removed_PLACEMARKER_OPEN * Fixed a blank white screen on the WPML Settings and Translation Dashboard pages in wp-admin.@removed_PLACEMARKER_CLOSED\n@removed_PLACEMARKER_OPEN * Fixed slow loading on the Menu Sync page.@removed_PLACEMARKER_CLOSED\n@removed_PLACEMARKER_OPEN Purchase@removed_PLACEMARKER_CLOSED\n@removed_PLACEMARKER_OPEN@removed_PLACEMARKER_CLOSED\n@removed_PLACEMARKER_OPEN Version 4.9.3@removed_PLACEMARKER_CLOSED\n@removed_PLACEMARKER_OPEN@removed_PLACEMARKER_CLOSED\n@removed_PLACEMARKER_OPEN May 6, 2026@removed_PLACEMARKER_CLOSED\n@removed_PLACEMARKER_OPEN@removed_PLACEMARKER_CLOSED\n@removed_PLACEMARKER_OPEN WPML 4.9.3 brings a wide range of fixes and improvements \u2014 from site migration and string translation to performance, admin usability, and editor compatibility.@removed_PLACEMARKER_CLOSED\n@removed_PLACEMARKER_OPEN@removed_PLACEMARKER_CLOSED\n@removed_PLACEMARKER_OPEN Features@removed_PLACEMARKER_CLOSED\n@removed_PLACEMARKER_OPEN@removed_PLACEMARKER_CLOSED\n@removed_PLACEMARKER_OPEN * New filter hook for developers to return translated user meta values@removed_PLACEMARKER_CLOSED\n@removed_PLACEMARKER_OPEN WPML now intercepts calls to get_user_meta() and returns the translated value for the current language. This is useful when displaying user profile fields (e.g. job title, bio) on multilingual sites. To enable this behaviour, add the following to your theme or plugin:@removed_PLACEMARKER_CLOSED\n@removed_PLACEMARKER_OPEN@removed_PLACEMARKER_CLOSED\n@removed_PLACEMARKER_OPEN add_filter( 'wpml_translate_get_user_meta', '__return_true' );@removed_PLACEMARKER_CLOSED\n@removed_PLACEMARKER_OPEN * New filter hook to globally disable all WPML email notifications@removed_PLACEMARKER_CLOSED\n@removed_PLACEMARKER_OPEN WPML sends automatic emails from several internal components (translation job assignments, overdue job reports, batch summaries, etc.). Previously there was no official way to suppress them \u2014 developers had to rely on workarounds inspecting the call stack. This new hook provides a single, supported way to disable all WPML emails at once:@removed_PLACEMARKER_CLOSED\n@removed_PLACEMARKER_OPEN@removed_PLACEMARKER_CLOSED\n@removed_PLACEMARKER_OPEN add_filter( 'wpml_send_email_notification', '__return_false' );@removed_PLACEMARKER_CLOSED\n@removed_PLACEMARKER_OPEN@removed_PLACEMARKER_CLOSED\n@removed_PLACEMARKER_OPEN Or selectively, based on email type and recipient:@removed_PLACEMARKER_CLOSED\n@removed_PLACEMARKER_OPEN@removed_PLACEMARKER_CLOSED\n@removed_PLACEMARKER_OPEN add_filter( 'wpml_send_email_notification', function( $send, $email_type, $recipient ) { if ( $email_type === 'overdue_jobs_report' ) { return false; } return $send; }, 10, 3 );@removed_PLACEMARKER_CLOSED\n@removed_PLACEMARKER_OPEN@removed_PLACEMARKER_CLOSED\n@removed_PLACEMARKER_OPEN This is particularly useful on staging or development environments where you want to prevent emails from reaching real users.@removed_PLACEMARKER_CLOSED\n@removed_PLACEMARKER_OPEN@removed_PLACEMARKER_CLOSED\n@removed_PLACEMARKER_OPEN Improvements@removed_PLACEMARKER_CLOSED\n@removed_PLACEMARKER_OPEN@removed_PLACEMARKER_CLOSED\n@removed_PLACEMARKER_OPEN Migration and Setup@removed_PLACEMARKER_CLOSED\n@removed_PLACEMARKER_OPEN@removed_PLACEMARKER_CLOSED\n@removed_PLACEMARKER_OPEN * Safer and More Reliable Migration Process@removed_PLACEMARKER_CLOSED\n@removed_PLACEMARKER_OPEN WPML now makes it safer and easier to move or copy a site from one domain to another. If something goes wrong during the migration, the wizard stays visible so users can retry, with clearer messages, better logging, and guidance on where to find the migration code. WPML also automatically fixes misconfigured alias domains, removes the unsafe \u201cMove\u201d option, and improves performance for sites using the Classic Translation Editor.@removed_PLACEMARKER_CLOSED\n@removed_PLACEMARKER_OPEN * Added a new option to reset the WPML Setup Wizard when it gets into a corrupted state.@removed_PLACEMARKER_CLOSED\n@removed_PLACEMARKER_OPEN * WPML now notifies the translation service when a site key is unregistered.@removed_PLACEMARKER_CLOSED\n@removed_PLACEMARKER_OPEN@removed_PLACEMARKER_CLOSED\n@removed_PLACEMARKER_OPEN Performance@removed_PLACEMARKER_CLOSED\n@removed_PLACEMARKER_OPEN@removed_PLACEMARKER_CLOSED\n@removed_PLACEMARKER_OPEN * Reduced redundant language filtering calls on sites with many ACF fields for better performance.@removed_PLACEMARKER_CLOSED\n@removed_PLACEMARKER_OPEN * Removed outdated WP Super Cache integration code that was no longer needed.@removed_PLACEMARKER_CLOSED\n@removed_PLACEMARKER_OPEN@removed_PLACEMARKER_CLOSED\n@removed_PLACEMARKER_OPEN Admin and User Experience@removed_PLACEMARKER_CLOSED\n@removed_PLACEMARKER_OPEN@removed_PLACEMARKER_CLOSED\n@removed_PLACEMARKER_OPEN * Unlocked WPML settings are now visually indicated in the admin interface.@removed_PLACEMARKER_CLOSED\n@removed_PLACEMARKER_OPEN * String Translation dashboard filters now show labels and default to Frontend strings.@removed_PLACEMARKER_CLOSED\n@removed_PLACEMARKER_OPEN * Clarified how Media Translation settings affect images and featured images.@removed_PLACEMARKER_CLOSED\n@removed_PLACEMARKER_OPEN * The notification about jobs that cannot be finished now links directly to the Translation Dashboard.@removed_PLACEMARKER_CLOSED\n@removed_PLACEMARKER_OPEN@removed_PLACEMARKER_CLOSED\n@removed_PLACEMARKER_OPEN Fixes@removed_PLACEMARKER_CLOSED\n@removed_PLACEMARKER_OPEN@removed_PLACEMARKER_CLOSED\n@removed_PLACEMARKER_OPEN * Some posts containing Base64-encoded data were not being sent for translation@removed_PLACEMARKER_CLOSED\n@removed_PLACEMARKER_OPEN In a previous release, we started blocking some posts because they contained Base64-encoded data, to prevent issues during the translation process. However, this also blocked some posts that could actually be translated, or where the Base64 content could simply be ignored. We\u2019ve updated how WPML handles Base64-encoded content \u2014 posts that can be safely translated are no longer blocked. A more precise validation approach is planned for an upcoming release.@removed_PLACEMARKER_CLOSED\n@removed_PLACEMARKER_OPEN * Fixed an issue with WPML blocks causing the Gutenberg editor to use an outdated rendering mode.@removed_PLACEMARKER_CLOSED\n@removed_PLACEMARKER_OPEN * Fixed an issue with Payments & Maintenance tab failing to load on sites with WordPress installed in a subdirectory.@removed_PLACEMARKER_CLOSED\n@removed_PLACEMARKER_OPEN * Fixed a PHP warning triggered when search queries contained array parameters.@removed_PLACEMARKER_CLOSED\n@removed_PLACEMARKER_OPEN * Fixed an issue with Full Site Editing category templates not applying to translated pages.@removed_PLACEMARKER_CLOSED\n@removed_PLACEMARKER_OPEN * Fixed an issue with translator assignment failing after a database restore or site migration.@removed_PLACEMARKER_CLOSED\n@removed_PLACEMARKER_OPEN * Fixed an issue with active subscriptions incorrectly shown as expired on the last day of the billing cycle.@removed_PLACEMARKER_CLOSED\n@removed_PLACEMARKER_OPEN * Fixed an issue with String Translation failing to load translation files due to domain name casing conflicts.@removed_PLACEMARKER_CLOSED\n@removed_PLACEMARKER_OPEN * Performance improvements when the Classic Translation Editor is selected.@removed_PLACEMARKER_CLOSED\n@removed_PLACEMARKER_OPEN * Fixed an issue with memory exhaustion errors on sites where user accounts had accented characters in their username.@removed_PLACEMARKER_CLOSED\n@removed_PLACEMARKER_OPEN * Fixed an issue with minimal custom field values being lost during the translation process.@removed_PLACEMARKER_CLOSED\n@removed_PLACEMARKER_OPEN * Fixed an issue with page preview slowness or timeouts on resource-limited servers when WPML is active.@removed_PLACEMARKER_CLOSED\n@removed_PLACEMARKER_OPEN * Fixed an issue with REST API requests using non-pretty URLs being incorrectly redirected.@removed_PLACEMARKER_CLOSED\n@removed_PLACEMARKER_OPEN * Fixed an issue with CSS changes in the WPML development project not being compiled correctly.@removed_PLACEMARKER_CLOSED\n@removed_PLACEMARKER_OPEN * Fixed a privacy policy link showing the original language page title instead of the translated one.@removed_PLACEMARKER_CLOSED\n@removed_PLACEMARKER_OPEN * Fixed an issue with the alias domain notice showing the same URL twice during site migrations.@removed_PLACEMARKER_CLOSED\n@removed_PLACEMARKER_OPEN * Fixed internal links using Cyrillic or non-Latin characters not being translated correctly.@removed_PLACEMARKER_CLOSED\n@removed_PLACEMARKER_OPEN * Fixed an issue with category and taxonomy terms not triggering the re-translation process correctly.@removed_PLACEMARKER_CLOSED\n@removed_PLACEMARKER_OPEN * Fixed an issue with browser language redirect not working correctly when a root page is set.@removed_PLACEMARKER_CLOSED\n@removed_PLACEMARKER_OPEN * Fixed an issue with the Classic Translation Editor causing PHP warnings and translation status loss on reload.@removed_PLACEMARKER_CLOSED\n@removed_PLACEMARKER_OPEN * Fixed an issue that was allowing posts to be accessed via the wrong category URL.@removed_PLACEMARKER_CLOSED\n@removed_PLACEMARKER_OPEN * Fixed an issue where String Translation file scanning was causing database corruption on large sites.@removed_PLACEMARKER_CLOSED\n@removed_PLACEMARKER_OPEN * Fixed an issue with the translation feedback button displaying incorrectly on Arabic (RTL) pages.@removed_PLACEMARKER_CLOSED\n@removed_PLACEMARKER_OPEN * Fixed a menu synchronization performance issue causing slowness or crashes on large multilingual sites.@removed_PLACEMARKER_CLOSED\n@removed_PLACEMARKER_OPEN * Fixed an issue with the Automatic Translation Payment & Maintenance tab failing to load on some sites.@removed_PLACEMARKER_CLOSED\n@removed_PLACEMARKER_OPEN * Fixed an issue with the development site banner failing the WCAG accessibility contrast requirements.@removed_PLACEMARKER_CLOSED\n@removed_PLACEMARKER_OPEN * Fixed an issue with internal links using www vs. non-www variants not being automatically translated.@removed_PLACEMARKER_CLOSED\n@removed_PLACEMARKER_OPEN * Fixed an issue with several WPML interface strings not being translated due to incorrect domain assignments.@removed_PLACEMARKER_CLOSED\n@removed_PLACEMARKER_OPEN * Fixed an issue with translated Elementor templates not displaying correctly on the frontend.@removed_PLACEMARKER_CLOSED\n@removed_PLACEMARKER_OPEN * Fixed an issue with Media Translation causing images to appear duplicated in the Media Library.@removed_PLACEMARKER_CLOSED\n@removed_PLACEMARKER_OPEN * Fixed an issue with the Save button remaining disabled when clearing translation context fields.@removed_PLACEMARKER_CLOSED\n@removed_PLACEMARKER_OPEN * Fixed an issue with website UUID mismatches causing translation service connectivity problems.@removed_PLACEMARKER_CLOSED\n@removed_PLACEMARKER_OPEN * Fixed WCAG accessibility issues with the legacy language switcher and SSO iframe.@removed_PLACEMARKER_CLOSED\n@removed_PLACEMARKER_OPEN * Fixed an issue with translator search causing heavy database load and CPU spikes.@removed_PLACEMARKER_CLOSED\n@removed_PLACEMARKER_OPEN * Fixed the word count calculation for translation jobs failing on Nginx/Kubernetes environments.@removed_PLACEMARKER_CLOSED\n@removed_PLACEMARKER_OPEN * Fixed a PHP 8.3 compatibility issue causing the Translation Dashboard to freeze.@removed_PLACEMARKER_CLOSED\n@removed_PLACEMARKER_OPEN * Fixed an issue with alias domain setting not being automatically removed when a domain mismatch is detected.@removed_PLACEMARKER_CLOSED\n@removed_PLACEMARKER_OPEN * Fixed a vulnerability in a WPML JavaScript file reported by SonarQube.@removed_PLACEMARKER_CLOSED\n@removed_PLACEMARKER_OPEN * Fixed an issue with class name causing issue in analytics code that could cause failures on Linux servers.@removed_PLACEMARKER_CLOSED\n@removed_PLACEMARKER_OPEN Purchase@removed_PLACEMARKER_CLOSED\n@removed_PLACEMARKER_OPEN@removed_PLACEMARKER_CLOSED\n@removed_PLACEMARKER_OPEN Version 4.9.2.1@removed_PLACEMARKER_CLOSED\n@removed_PLACEMARKER_OPEN@removed_PLACEMARKER_CLOSED\n@removed_PLACEMARKER_OPEN April 21, 2026@removed_PLACEMARKER_CLOSED\n@removed_PLACEMARKER_OPEN@removed_PLACEMARKER_CLOSED\n@removed_PLACEMARKER_OPEN Fixes@removed_PLACEMARKER_CLOSED\n@removed_PLACEMARKER_OPEN@removed_PLACEMARKER_CLOSED\n@removed_PLACEMARKER_OPEN * Fixed a security issue in WPML\u2019s Translation Feedback feature. Heads-up: we\u2019re retiring this feature in the next major WPML release because the usage has been very low.@removed_PLACEMARKER_CLOSED\n@removed_PLACEMARKER_OPEN * Fixed an issue where WooCommerce customers could experience a slow checkout and see a false payment error, even though their order had been successfully placed and their payment had gone through.@removed_PLACEMARKER_CLOSED\n@removed_PLACEMARKER_OPEN * Fixed an issue where users could see a false error when copying their site during the migration process, even though the operation had completed successfully in the background.@removed_PLACEMARKER_CLOSED\n@removed_PLACEMARKER_OPEN Purchase@removed_PLACEMARKER_CLOSED", "markup_text_links_to_html_links": false, "restock": {}, "notification_timestamp": 1784567684.1868193, "prev_snapshot": " Version 4.9.5\n\n June 10, 2026\n\n Compatibility\n\n * WordPress 7.0 \u2014 Fixed the Classic Translation Editor field layout, the post-language-change modal, and the language switcher block so they render correctly on WordPress 7.0.\n * PHP 8.4 and 8.5 \u2014 Cleaned up deprecation notices and silent warnings so sites on the latest PHP versions stay clean in their error logs.\n * Elementor \u2014 Strings sent to the Advanced Translation Editor no longer include inactive fields.\n\n Moving and re-registering a site\n\n * When you move a site to a new domain, WPML now applies the Copy with migration code option automatically, dismisses stale migration notices the moment the URL changes, and shows the migration banner as soon as the alias domain is reset.\n * Restored the explanation shown when the alias domain is automatically reset, and re-enabled the manual alias-domain entry on the Troubleshooting page so site registration is reachable again on sites upgrading from 4.9.4.\n * Stale cloned site flags are now cleared on upgrade so a partially-completed migration wizard no longer blocks subsequent runs.\n * Redesigned the refund / unregister / replace-site-key notice as a clearer three-card layout, making it easier to remove WPML from a site you no longer use.\n\n Advanced Translation Editor\n\n * Slug changes made in the Advanced Translation Editor are now correctly saved back to the translated post.\n * The lock-contention response is now a proper HTTP 423 status instead of HTTP 500, so retries from the translation service work correctly.\n * Translation jobs are now sent in priority order \u2014 items at the top of your queue are translated first.\n * Translations downloaded from the Advanced Translation Editor now appear in the language switcher immediately, without needing a cache flush.\n * Translation sync error screens now include support instructions and direct links to the relevant help articles.\n * The Translation Management job log was rebuilt on a new storage backend with much richer logging, making it easier to diagnose translation issues.\n\n Translation Management dashboard\n\n * The post edit screen is significantly faster on sites with orphan posts created in secondary languages \u2014 the per-render orphan-translations check has been replaced with a much lighter query, so editing posts in Gutenberg no longer pauses on large sites.\n * Removed unused legacy tracking code and a duplicate JavaScript global from the front of the admin.\n\n URLs, slugs, language switcher\n\n * Plain-URL to permalink conversion now preserves the port number.\n * The translated URL is generated the first time when translating with the native WordPress editor \u2014 no more translate twice to see the slug.\n * Fixed translation of links at the bottom of toolset archive pages.\n * An empty translation ID no longer accidentally creates a stray post during in-progress Classic Editor translations.\n * Private posts no longer get redirected to a media post when they share a slug.\n * Private posts now appear on the secondary-language blog page.\n * Future-status posts now show the language switcher again.\n * Language switcher block: fixed flag size in the preview area, and added explicit width and height attributes on flag images for better Core Web Vitals.\n\n Menus, caches, and miscellaneous\n\n * Sites running an object cache no longer lose translation menu items after Sync Menus \u2014 the term cache is now invalidated correctly.\n * ACF Gallery Block: duplicating a page now duplicates all images, not just the first one.\n * Cleared a false-positive SQL-injection warning from third-party security scanners.\n Purchase\n\n Version 4.9.4\n\n May 19, 2026\n\n WPML 4.9.4 brings full compatibility with WordPress 7.0, addressing several changes to the block editor and core functionality introduced in this release.\n\n WordPress 7.0 Compatibility\n\n * The block editor now stays in modern iframe mode on WPML-enabled sites. WordPress 7.0 makes the fully iframe-based editor the new standard. On sites running an older version of WPML, the editor would silently revert to legacy mode whenever WPML was active. The Language Switcher and Navigation Language Switcher blocks have been updated to fully support Block API version 3, keeping the editor in modern mode as expected.\n * Navigation Language Switcher \u201copen on click\u201d now works correctly on WordPress 7.0. WordPress 7.0 changed how the Navigation block stores this setting. WPML now reads the updated value correctly, so submenus respect the behaviour you configured.\n * New WordPress 7.0 block content is now translatable. Three elements introduced or restructured in WordPress 7.0 were not appearing in the WPML translation editor. All three are now fully translatable: Icon block accessibility labels (aria-label), Navigation overlay close button text, and Tab block labels.\n * Fixed block editor warnings and styling issues on WordPress 7.0. WPML\u2019s internal code for registering block assets has been updated to use current WordPress APIs. Block descriptions are now plain strings as WordPress 7.0 requires, and CSS loads correctly inside the iframe editor. All changes are backward compatible to WordPress 6.3.\n * Fixed PHP deprecation notices in server logs. A PHP function deprecated in WordPress 7.0 has been replaced with the modern equivalent throughout WPML\u2019s codebase.\n\n Other Fixes\n\n * Fixed a blank white screen on the WPML Settings and Translation Dashboard pages in wp-admin.\n * Fixed slow loading on the Menu Sync page.\n Purchase\n\n Version 4.9.3\n\n May 6, 2026\n\n WPML 4.9.3 brings a wide range of fixes and improvements \u2014 from site migration and string translation to performance, admin usability, and editor compatibility.\n\n Features\n\n * New filter hook for developers to return translated user meta values\n WPML now intercepts calls to get_user_meta() and returns the translated value for the current language. This is useful when displaying user profile fields (e.g. job title, bio) on multilingual sites. To enable this behaviour, add the following to your theme or plugin:\n\n add_filter( 'wpml_translate_get_user_meta', '__return_true' );\n * New filter hook to globally disable all WPML email notifications\n WPML sends automatic emails from several internal components (translation job assignments, overdue job reports, batch summaries, etc.). Previously there was no official way to suppress them \u2014 developers had to rely on workarounds inspecting the call stack. This new hook provides a single, supported way to disable all WPML emails at once:\n\n add_filter( 'wpml_send_email_notification', '__return_false' );\n\n Or selectively, based on email type and recipient:\n\n add_filter( 'wpml_send_email_notification', function( $send, $email_type, $recipient ) { if ( $email_type === 'overdue_jobs_report' ) { return false; } return $send; }, 10, 3 );\n\n This is particularly useful on staging or development environments where you want to prevent emails from reaching real users.\n\n Improvements\n\n Migration and Setup\n\n * Safer and More Reliable Migration Process\n WPML now makes it safer and easier to move or copy a site from one domain to another. If something goes wrong during the migration, the wizard stays visible so users can retry, with clearer messages, better logging, and guidance on where to find the migration code. WPML also automatically fixes misconfigured alias domains, removes the unsafe \u201cMove\u201d option, and improves performance for sites using the Classic Translation Editor.\n * Added a new option to reset the WPML Setup Wizard when it gets into a corrupted state.\n * WPML now notifies the translation service when a site key is unregistered.\n\n Performance\n\n * Reduced redundant language filtering calls on sites with many ACF fields for better performance.\n * Removed outdated WP Super Cache integration code that was no longer needed.\n\n Admin and User Experience\n\n * Unlocked WPML settings are now visually indicated in the admin interface.\n * String Translation dashboard filters now show labels and default to Frontend strings.\n * Clarified how Media Translation settings affect images and featured images.\n * The notification about jobs that cannot be finished now links directly to the Translation Dashboard.\n\n Fixes\n\n * Some posts containing Base64-encoded data were not being sent for translation\n In a previous release, we started blocking some posts because they contained Base64-encoded data, to prevent issues during the translation process. However, this also blocked some posts that could actually be translated, or where the Base64 content could simply be ignored. We\u2019ve updated how WPML handles Base64-encoded content \u2014 posts that can be safely translated are no longer blocked. A more precise validation approach is planned for an upcoming release.\n * Fixed an issue with WPML blocks causing the Gutenberg editor to use an outdated rendering mode.\n * Fixed an issue with Payments & Maintenance tab failing to load on sites with WordPress installed in a subdirectory.\n * Fixed a PHP warning triggered when search queries contained array parameters.\n * Fixed an issue with Full Site Editing category templates not applying to translated pages.\n * Fixed an issue with translator assignment failing after a database restore or site migration.\n * Fixed an issue with active subscriptions incorrectly shown as expired on the last day of the billing cycle.\n * Fixed an issue with String Translation failing to load translation files due to domain name casing conflicts.\n * Performance improvements when the Classic Translation Editor is selected.\n * Fixed an issue with memory exhaustion errors on sites where user accounts had accented characters in their username.\n * Fixed an issue with minimal custom field values being lost during the translation process.\n * Fixed an issue with page preview slowness or timeouts on resource-limited servers when WPML is active.\n * Fixed an issue with REST API requests using non-pretty URLs being incorrectly redirected.\n * Fixed an issue with CSS changes in the WPML development project not being compiled correctly.\n * Fixed a privacy policy link showing the original language page title instead of the translated one.\n * Fixed an issue with the alias domain notice showing the same URL twice during site migrations.\n * Fixed internal links using Cyrillic or non-Latin characters not being translated correctly.\n * Fixed an issue with category and taxonomy terms not triggering the re-translation process correctly.\n * Fixed an issue with browser language redirect not working correctly when a root page is set.\n * Fixed an issue with the Classic Translation Editor causing PHP warnings and translation status loss on reload.\n * Fixed an issue that was allowing posts to be accessed via the wrong category URL.\n * Fixed an issue where String Translation file scanning was causing database corruption on large sites.\n * Fixed an issue with the translation feedback button displaying incorrectly on Arabic (RTL) pages.\n * Fixed a menu synchronization performance issue causing slowness or crashes on large multilingual sites.\n * Fixed an issue with the Automatic Translation Payment & Maintenance tab failing to load on some sites.\n * Fixed an issue with the development site banner failing the WCAG accessibility contrast requirements.\n * Fixed an issue with internal links using www vs. non-www variants not being automatically translated.\n * Fixed an issue with several WPML interface strings not being translated due to incorrect domain assignments.\n * Fixed an issue with translated Elementor templates not displaying correctly on the frontend.\n * Fixed an issue with Media Translation causing images to appear duplicated in the Media Library.\n * Fixed an issue with the Save button remaining disabled when clearing translation context fields.\n * Fixed an issue with website UUID mismatches causing translation service connectivity problems.\n * Fixed WCAG accessibility issues with the legacy language switcher and SSO iframe.\n * Fixed an issue with translator search causing heavy database load and CPU spikes.\n * Fixed the word count calculation for translation jobs failing on Nginx/Kubernetes environments.\n * Fixed a PHP 8.3 compatibility issue causing the Translation Dashboard to freeze.\n * Fixed an issue with alias domain setting not being automatically removed when a domain mismatch is detected.\n * Fixed a vulnerability in a WPML JavaScript file reported by SonarQube.\n * Fixed an issue with class name causing issue in analytics code that could cause failures on Linux servers.\n Purchase\n\n Version 4.9.2.1\n\n April 21, 2026\n\n Fixes\n\n * Fixed a security issue in WPML\u2019s Translation Feedback feature. Heads-up: we\u2019re retiring this feature in the next major WPML release because the usage has been very low.\n * Fixed an issue where WooCommerce customers could experience a slow checkout and see a false payment error, even though their order had been successfully placed and their payment had gone through.\n * Fixed an issue where users could see a false error when copying their site during the migration process, even though the operation had completed successfully in the background.\n Purchase\n", "preview_url": "https://changedetection.benfou.fr/preview/83594c1a-d307-48ee-a2dd-1586170fc4f4", "screenshot": null, "timestamp_from": null, "timestamp_to": null, "triggered_text": "", "llm_summary": null, "llm_intent": null, "uuid": "83594c1a-d307-48ee-a2dd-1586170fc4f4", "watch_mime_type": "text/html; charset=utf-8", "watch_tag": "Technologie", "watch_title": "Update WPML", "watch_url": "https://wpml.org/download/wpml-multilingual-cms/?section=changelog", "watch_uuid": "83594c1a-d307-48ee-a2dd-1586170fc4f4", "notification_urls": ["jsons://n8n.benfou.fr/webhook/c8bb38ef-b9d8-45d0-837d-c239411db01b/?title={{ watch_title|urlencode }}"], "notification_title": "Changement d\u00e9tect\u00e9 - {{watch_title}}", "notification_body": "{{watch_url}} a chang\u00e9e.\r\n---\r\n{{diff}}\r\n---\r\n", "notification_format": "htmlcolor", "_llm_result": null, "_llm_intent": "", "_llm_change_summary": "", "edit_url": "https://changedetection.benfou.fr/edit/83594c1a-d307-48ee-a2dd-1586170fc4f4"}}]
Sun Jul 19 07:13:45 2026 - SENDING - [{"title": "Changement d\u00e9tect\u00e9 - Update OpenClaw", "body": "https://github.com/openclaw/openclaw/releases a chang\u00e9e. \n--- \n      * openclaw 2026.7.2-beta.22026.7.2-beta.3 \n---", "url": "jsons://n8n.benfou.fr/webhook/c8bb38ef-b9d8-45d0-837d-c239411db01b/?title=Update%20OpenClaw&avatar_url=https://raw.githubusercontent.com/dgtlmoon/changedetection.io/master/changedetectionio/static/images/avatar-256x256.png&format=html", "original_context": {"base_url": "https://changedetection.benfou.fr/", "change_datetime": "2026-07-19 07:13:03 CEST", "current_snapshot": " * openclaw 2026.7.2-beta.3", "diff": " * openclaw @removed_PLACEMARKER_OPEN2026.7.2-beta.2@removed_PLACEMARKER_CLOSED@added_PLACEMARKER_OPEN2026.7.2-beta.3@added_PLACEMARKER_CLOSED", "diff_clean": "", "diff_added": "", "diff_added_clean": "", "diff_full": "", "diff_full_clean": "", "diff_patch": "", "diff_removed": "", "diff_removed_clean": "", "diff_changed_from": "", "diff_changed_to": "", "diff_url": "https://changedetection.benfou.fr/diff/27f8a5ca-cd09-4e82-8b3c-5eb340fe0bfc", "raw_diff": " * openclaw @removed_PLACEMARKER_OPEN2026.7.2-beta.2@removed_PLACEMARKER_CLOSED@added_PLACEMARKER_OPEN2026.7.2-beta.3@added_PLACEMARKER_CLOSED", "markup_text_links_to_html_links": false, "restock": {}, "notification_timestamp": 1784438025.1966949, "prev_snapshot": " * openclaw 2026.7.2-beta.2", "preview_url": "https://changedetection.benfou.fr/preview/27f8a5ca-cd09-4e82-8b3c-5eb340fe0bfc", "screenshot": null, "timestamp_from": null, "timestamp_to": null, "triggered_text": "", "llm_summary": null, "llm_intent": null, "uuid": "27f8a5ca-cd09-4e82-8b3c-5eb340fe0bfc", "watch_mime_type": "text/html; charset=utf-8", "watch_tag": "Technologie", "watch_title": "Update OpenClaw", "watch_url": "https://github.com/openclaw/openclaw/releases", "watch_uuid": "27f8a5ca-cd09-4e82-8b3c-5eb340fe0bfc", "notification_urls": ["jsons://n8n.benfou.fr/webhook/c8bb38ef-b9d8-45d0-837d-c239411db01b/?title={{ watch_title|urlencode }}"], "notification_title": "Changement d\u00e9tect\u00e9 - {{watch_title}}", "notification_body": "{{watch_url}} a chang\u00e9e.\r\n---\r\n{{diff}}\r\n---\r\n", "notification_format": "htmlcolor", "_llm_result": null, "_llm_intent": "", "_llm_change_summary": "", "edit_url": "https://changedetection.benfou.fr/edit/27f8a5ca-cd09-4e82-8b3c-5eb340fe0bfc"}}]
Fri Jul 17 19:13:33 2026 - SENDING - [{"title": "Changement d\u00e9tect\u00e9 - Update OpenClaw", "body": "https://github.com/openclaw/openclaw/releases a chang\u00e9e. \n--- \n      * openclaw 2026.7.2-beta.12026.7.2-beta.2 \n---", "url": "jsons://n8n.benfou.fr/webhook/c8bb38ef-b9d8-45d0-837d-c239411db01b/?title=Update%20OpenClaw&avatar_url=https://raw.githubusercontent.com/dgtlmoon/changedetection.io/master/changedetectionio/static/images/avatar-256x256.png&format=html", "original_context": {"base_url": "https://changedetection.benfou.fr/", "change_datetime": "2026-07-17 19:12:59 CEST", "current_snapshot": " * openclaw 2026.7.2-beta.2", "diff": " * openclaw @removed_PLACEMARKER_OPEN2026.7.2-beta.1@removed_PLACEMARKER_CLOSED@added_PLACEMARKER_OPEN2026.7.2-beta.2@added_PLACEMARKER_CLOSED", "diff_clean": "", "diff_added": "", "diff_added_clean": "", "diff_full": "", "diff_full_clean": "", "diff_patch": "", "diff_removed": "", "diff_removed_clean": "", "diff_changed_from": "", "diff_changed_to": "", "diff_url": "https://changedetection.benfou.fr/diff/27f8a5ca-cd09-4e82-8b3c-5eb340fe0bfc", "raw_diff": " * openclaw @removed_PLACEMARKER_OPEN2026.7.2-beta.1@removed_PLACEMARKER_CLOSED@added_PLACEMARKER_OPEN2026.7.2-beta.2@added_PLACEMARKER_CLOSED", "markup_text_links_to_html_links": false, "restock": {}, "notification_timestamp": 1784308412.9449768, "prev_snapshot": " * openclaw 2026.7.2-beta.1", "preview_url": "https://changedetection.benfou.fr/preview/27f8a5ca-cd09-4e82-8b3c-5eb340fe0bfc", "screenshot": null, "timestamp_from": null, "timestamp_to": null, "triggered_text": "", "llm_summary": null, "llm_intent": null, "uuid": "27f8a5ca-cd09-4e82-8b3c-5eb340fe0bfc", "watch_mime_type": "text/html; charset=utf-8", "watch_tag": "Technologie", "watch_title": "Update OpenClaw", "watch_url": "https://github.com/openclaw/openclaw/releases", "watch_uuid": "27f8a5ca-cd09-4e82-8b3c-5eb340fe0bfc", "notification_urls": ["jsons://n8n.benfou.fr/webhook/c8bb38ef-b9d8-45d0-837d-c239411db01b/?title={{ watch_title|urlencode }}"], "notification_title": "Changement d\u00e9tect\u00e9 - {{watch_title}}", "notification_body": "{{watch_url}} a chang\u00e9e.\r\n---\r\n{{diff}}\r\n---\r\n", "notification_format": "htmlcolor", "_llm_result": null, "_llm_intent": "", "_llm_change_summary": "", "edit_url": "https://changedetection.benfou.fr/edit/27f8a5ca-cd09-4e82-8b3c-5eb340fe0bfc"}}]
Fri Jul 17 19:12:54 2026 - SENDING - [{"title": "Changement d\u00e9tect\u00e9 - API Claude", "body": "https://platform.claude.com/docs/en/about-claude/pricing a chang\u00e9e. \n--- \n                                * Additional usage beyond 1,550 hours is billed at $0.05 per hour, per container \n                                * If files are included in the request, execution time is billed even if the tool is not invoked, because files are preloaded onto the container \n                                * Additional usage beyond 1,550 hours is billed at $0.05 USD per hour, per container \n                                * If files are included in the request, execution time is billed even if the tool is not called, because files are preloaded onto the container \n                                System prompt overhead: The computer use beta adds 466-499466\u2013499 tokens to the system prompt \n              Search... \n              Search \n---", "url": "jsons://n8n.benfou.fr/webhook/c8bb38ef-b9d8-45d0-837d-c239411db01b/?title=API%20Claude&avatar_url=https://raw.githubusercontent.com/dgtlmoon/changedetection.io/master/changedetectionio/static/images/avatar-256x256.png&format=html", "original_context": {"base_url": "https://changedetection.benfou.fr/", "change_datetime": "2026-07-17 19:12:19 CEST", "current_snapshot": " Search...\n Search", "diff": "@changed_PLACEMARKER_OPEN * Additional usage beyond 1,550 hours is billed at $0.05 per hour, per container@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * If files are included in the request, execution time is billed even if the tool is not invoked, because files are preloaded onto the container@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Additional usage beyond 1,550 hours is billed at $0.05 USD per hour, per container@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * If files are included in the request, execution time is billed even if the tool is not called, because files are preloaded onto the container@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN System prompt overhead: The computer use beta adds 466-499466\u2013499 tokens to the system prompt@changed_PLACEMARKER_CLOSED\n@changed_into_PLACEMARKER_OPEN Search...@changed_into_PLACEMARKER_CLOSED\n@changed_into_PLACEMARKER_OPEN Search@changed_into_PLACEMARKER_CLOSED", "diff_clean": "", "diff_added": "", "diff_added_clean": "", "diff_full": "", "diff_full_clean": "", "diff_patch": "", "diff_removed": "", "diff_removed_clean": "", "diff_changed_from": "", "diff_changed_to": "", "diff_url": "https://changedetection.benfou.fr/diff/f10610bf-32e8-4762-a1c2-e9481624471d", "raw_diff": "@changed_PLACEMARKER_OPEN * Additional usage beyond 1,550 hours is billed at $0.05 per hour, per container@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * If files are included in the request, execution time is billed even if the tool is not invoked, because files are preloaded onto the container@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Additional usage beyond 1,550 hours is billed at $0.05 USD per hour, per container@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * If files are included in the request, execution time is billed even if the tool is not called, because files are preloaded onto the container@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN System prompt overhead: The computer use beta adds 466-499466\u2013499 tokens to the system prompt@changed_PLACEMARKER_CLOSED\n@changed_into_PLACEMARKER_OPEN Search...@changed_into_PLACEMARKER_CLOSED\n@changed_into_PLACEMARKER_OPEN Search@changed_into_PLACEMARKER_CLOSED", "markup_text_links_to_html_links": false, "restock": {}, "notification_timestamp": 1784308374.2406528, "prev_snapshot": " * Additional usage beyond 1,550 hours is billed at $0.05 per hour, per container\n * If files are included in the request, execution time is billed even if the tool is not invoked, because files are preloaded onto the container\n * Additional usage beyond 1,550 hours is billed at $0.05 USD per hour, per container\n * If files are included in the request, execution time is billed even if the tool is not called, because files are preloaded onto the container\n System prompt overhead: The computer use beta adds 466-499466\u2013499 tokens to the system prompt", "preview_url": "https://changedetection.benfou.fr/preview/f10610bf-32e8-4762-a1c2-e9481624471d", "screenshot": null, "timestamp_from": null, "timestamp_to": null, "triggered_text": "", "llm_summary": null, "llm_intent": null, "uuid": "f10610bf-32e8-4762-a1c2-e9481624471d", "watch_mime_type": "text/html; charset=utf-8", "watch_tag": "IA", "watch_title": "API Claude", "watch_url": "https://platform.claude.com/docs/en/about-claude/pricing", "watch_uuid": "f10610bf-32e8-4762-a1c2-e9481624471d", "notification_urls": ["jsons://n8n.benfou.fr/webhook/c8bb38ef-b9d8-45d0-837d-c239411db01b/?title={{ watch_title|urlencode }}"], "notification_title": "Changement d\u00e9tect\u00e9 - {{watch_title}}", "notification_body": "{{watch_url}} a chang\u00e9e.\r\n---\r\n{{diff}}\r\n---\r\n", "notification_format": "htmlcolor", "_llm_result": null, "_llm_intent": "", "_llm_change_summary": "", "edit_url": "https://changedetection.benfou.fr/edit/f10610bf-32e8-4762-a1c2-e9481624471d"}}]
Fri Jul 17 19:12:42 2026 - SENDING - [{"title": "Changement d\u00e9tect\u00e9 - API OpenIA", "body": "https://developers.openai.com/api/docs/changelog a chang\u00e9e. \n--- \n            * Codex Micro \n            * Projects, chats,Projects and taskschats \n            * Cloud FAQ \n                + Write vulnerability reports \n                + Propose security hardening \n            * ChatGPT Work Adminmode admin FAQ \n---", "url": "jsons://n8n.benfou.fr/webhook/c8bb38ef-b9d8-45d0-837d-c239411db01b/?title=API%20OpenIA&avatar_url=https://raw.githubusercontent.com/dgtlmoon/changedetection.io/master/changedetectionio/static/images/avatar-256x256.png&format=html", "original_context": {"base_url": "https://changedetection.benfou.fr/", "change_datetime": "2026-07-17 19:12:16 CEST", "current_snapshot": " * Projects, chats,Projects and taskschats\n * Cloud FAQ\n + Write vulnerability reports\n + Propose security hardening\n * ChatGPT Work Adminmode admin FAQ", "diff": "@changed_PLACEMARKER_OPEN * Codex Micro@changed_PLACEMARKER_CLOSED\n@changed_into_PLACEMARKER_OPEN * Projects, chats,Projects and taskschats@changed_into_PLACEMARKER_CLOSED\n@changed_into_PLACEMARKER_OPEN * Cloud FAQ@changed_into_PLACEMARKER_CLOSED\n@changed_into_PLACEMARKER_OPEN + Write vulnerability reports@changed_into_PLACEMARKER_CLOSED\n@changed_into_PLACEMARKER_OPEN + Propose security hardening@changed_into_PLACEMARKER_CLOSED\n@changed_into_PLACEMARKER_OPEN * ChatGPT Work Adminmode admin FAQ@changed_into_PLACEMARKER_CLOSED", "diff_clean": "", "diff_added": "", "diff_added_clean": "", "diff_full": "", "diff_full_clean": "", "diff_patch": "", "diff_removed": "", "diff_removed_clean": "", "diff_changed_from": "", "diff_changed_to": "", "diff_url": "https://changedetection.benfou.fr/diff/40d1d040-eab5-44c8-a830-96ce77c471dd", "raw_diff": "@changed_PLACEMARKER_OPEN * Codex Micro@changed_PLACEMARKER_CLOSED\n@changed_into_PLACEMARKER_OPEN * Projects, chats,Projects and taskschats@changed_into_PLACEMARKER_CLOSED\n@changed_into_PLACEMARKER_OPEN * Cloud FAQ@changed_into_PLACEMARKER_CLOSED\n@changed_into_PLACEMARKER_OPEN + Write vulnerability reports@changed_into_PLACEMARKER_CLOSED\n@changed_into_PLACEMARKER_OPEN + Propose security hardening@changed_into_PLACEMARKER_CLOSED\n@changed_into_PLACEMARKER_OPEN * ChatGPT Work Adminmode admin FAQ@changed_into_PLACEMARKER_CLOSED", "markup_text_links_to_html_links": false, "restock": {}, "notification_timestamp": 1784308361.233129, "prev_snapshot": " * Codex Micro", "preview_url": "https://changedetection.benfou.fr/preview/40d1d040-eab5-44c8-a830-96ce77c471dd", "screenshot": null, "timestamp_from": null, "timestamp_to": null, "triggered_text": "", "llm_summary": null, "llm_intent": null, "uuid": "40d1d040-eab5-44c8-a830-96ce77c471dd", "watch_mime_type": "text/html; charset=utf-8", "watch_tag": "IA", "watch_title": "API OpenIA", "watch_url": "https://developers.openai.com/api/docs/changelog", "watch_uuid": "40d1d040-eab5-44c8-a830-96ce77c471dd", "notification_urls": ["jsons://n8n.benfou.fr/webhook/c8bb38ef-b9d8-45d0-837d-c239411db01b/?title={{ watch_title|urlencode }}"], "notification_title": "Changement d\u00e9tect\u00e9 - {{watch_title}}", "notification_body": "{{watch_url}} a chang\u00e9e.\r\n---\r\n{{diff}}\r\n---\r\n", "notification_format": "htmlcolor", "_llm_result": null, "_llm_intent": "", "_llm_change_summary": "", "edit_url": "https://changedetection.benfou.fr/edit/40d1d040-eab5-44c8-a830-96ce77c471dd"}}]
Fri Jul 17 01:12:49 2026 - SENDING - [{"title": "Changement d\u00e9tect\u00e9 - API Claude", "body": "https://platform.claude.com/docs/en/about-claude/pricing a chang\u00e9e. \n--- \n                                * If files are included in the request, execution time is billed even if the tool is not invoked, due tobecause files beingare preloaded onto the container \n                                * Additional usage beyond 1,550 hours is billed at $0.05 per hour, per container \n                                * If files are included in the request, execution time is billed even if the tool is not invoked, because files are preloaded onto the container \n                                * Additional usage beyond 1,550 hours is billed at $0.05 USD per hour, per container \n                                * If files are included in the request, execution time is billed even if the tool is not called, because files are preloaded onto the container \n                                System prompt overhead: The computer use beta adds 466-499466\u2013499 tokens to the system prompt \n---", "url": "jsons://n8n.benfou.fr/webhook/c8bb38ef-b9d8-45d0-837d-c239411db01b/?title=API%20Claude&avatar_url=https://raw.githubusercontent.com/dgtlmoon/changedetection.io/master/changedetectionio/static/images/avatar-256x256.png&format=html", "original_context": {"base_url": "https://changedetection.benfou.fr/", "change_datetime": "2026-07-17 01:12:19 CEST", "current_snapshot": " * Additional usage beyond 1,550 hours is billed at $0.05 per hour, per container\n * If files are included in the request, execution time is billed even if the tool is not invoked, because files are preloaded onto the container\n * Additional usage beyond 1,550 hours is billed at $0.05 USD per hour, per container\n * If files are included in the request, execution time is billed even if the tool is not called, because files are preloaded onto the container\n System prompt overhead: The computer use beta adds 466-499466\u2013499 tokens to the system prompt", "diff": "@changed_PLACEMARKER_OPEN * If files are included in the request, execution time is billed even if the tool is not invoked, due tobecause files beingare preloaded onto the container@changed_PLACEMARKER_CLOSED\n@changed_into_PLACEMARKER_OPEN * Additional usage beyond 1,550 hours is billed at $0.05 per hour, per container@changed_into_PLACEMARKER_CLOSED\n@changed_into_PLACEMARKER_OPEN * If files are included in the request, execution time is billed even if the tool is not invoked, because files are preloaded onto the container@changed_into_PLACEMARKER_CLOSED\n@changed_into_PLACEMARKER_OPEN * Additional usage beyond 1,550 hours is billed at $0.05 USD per hour, per container@changed_into_PLACEMARKER_CLOSED\n@changed_into_PLACEMARKER_OPEN * If files are included in the request, execution time is billed even if the tool is not called, because files are preloaded onto the container@changed_into_PLACEMARKER_CLOSED\n@changed_into_PLACEMARKER_OPEN System prompt overhead: The computer use beta adds 466-499466\u2013499 tokens to the system prompt@changed_into_PLACEMARKER_CLOSED", "diff_clean": "", "diff_added": "", "diff_added_clean": "", "diff_full": "", "diff_full_clean": "", "diff_patch": "", "diff_removed": "", "diff_removed_clean": "", "diff_changed_from": "", "diff_changed_to": "", "diff_url": "https://changedetection.benfou.fr/diff/f10610bf-32e8-4762-a1c2-e9481624471d", "raw_diff": "@changed_PLACEMARKER_OPEN * If files are included in the request, execution time is billed even if the tool is not invoked, due tobecause files beingare preloaded onto the container@changed_PLACEMARKER_CLOSED\n@changed_into_PLACEMARKER_OPEN * Additional usage beyond 1,550 hours is billed at $0.05 per hour, per container@changed_into_PLACEMARKER_CLOSED\n@changed_into_PLACEMARKER_OPEN * If files are included in the request, execution time is billed even if the tool is not invoked, because files are preloaded onto the container@changed_into_PLACEMARKER_CLOSED\n@changed_into_PLACEMARKER_OPEN * Additional usage beyond 1,550 hours is billed at $0.05 USD per hour, per container@changed_into_PLACEMARKER_CLOSED\n@changed_into_PLACEMARKER_OPEN * If files are included in the request, execution time is billed even if the tool is not called, because files are preloaded onto the container@changed_into_PLACEMARKER_CLOSED\n@changed_into_PLACEMARKER_OPEN System prompt overhead: The computer use beta adds 466-499466\u2013499 tokens to the system prompt@changed_into_PLACEMARKER_CLOSED", "markup_text_links_to_html_links": false, "restock": {}, "notification_timestamp": 1784243568.563425, "prev_snapshot": " * If files are included in the request, execution time is billed even if the tool is not invoked, due tobecause files beingare preloaded onto the container", "preview_url": "https://changedetection.benfou.fr/preview/f10610bf-32e8-4762-a1c2-e9481624471d", "screenshot": null, "timestamp_from": null, "timestamp_to": null, "triggered_text": "", "llm_summary": null, "llm_intent": null, "uuid": "f10610bf-32e8-4762-a1c2-e9481624471d", "watch_mime_type": "text/html; charset=utf-8", "watch_tag": "IA", "watch_title": "API Claude", "watch_url": "https://platform.claude.com/docs/en/about-claude/pricing", "watch_uuid": "f10610bf-32e8-4762-a1c2-e9481624471d", "notification_urls": ["jsons://n8n.benfou.fr/webhook/c8bb38ef-b9d8-45d0-837d-c239411db01b/?title={{ watch_title|urlencode }}"], "notification_title": "Changement d\u00e9tect\u00e9 - {{watch_title}}", "notification_body": "{{watch_url}} a chang\u00e9e.\r\n---\r\n{{diff}}\r\n---\r\n", "notification_format": "htmlcolor", "_llm_result": null, "_llm_intent": "", "_llm_change_summary": "", "edit_url": "https://changedetection.benfou.fr/edit/f10610bf-32e8-4762-a1c2-e9481624471d"}}]
Thu Jul 16 07:13:52 2026 - SENDING - [{"title": "Changement d\u00e9tect\u00e9 - AquaPurr 4 - Battery powered cat water fountain - Never needs cleaning", "body": "https://aquapurr.com/products/aquapurr-5-refurbished a chang\u00e9e. \n--- \n      Sold Out \nIn Stock: True - Price: 99.0 \n---", "url": "jsons://n8n.benfou.fr/webhook/c8bb38ef-b9d8-45d0-837d-c239411db01b/?title=AquaPurr%204%20-%20Battery%20powered%20cat%20water%20fountain%20-%20Never%20needs%20cleaning&avatar_url=https://raw.githubusercontent.com/dgtlmoon/changedetection.io/master/changedetectionio/static/images/avatar-256x256.png&format=html", "original_context": {"base_url": "https://changedetection.benfou.fr/", "change_datetime": "2026-07-16 07:13:28 CEST", "current_snapshot": "In Stock: True - Price: 99.0", "diff": "@changed_PLACEMARKER_OPEN Sold Out@changed_PLACEMARKER_CLOSED\n@changed_into_PLACEMARKER_OPENIn Stock: True - Price: 99.0@changed_into_PLACEMARKER_CLOSED", "diff_clean": "", "diff_added": "", "diff_added_clean": "", "diff_full": "", "diff_full_clean": "", "diff_patch": "", "diff_removed": "", "diff_removed_clean": "", "diff_changed_from": "", "diff_changed_to": "", "diff_url": "https://changedetection.benfou.fr/diff/c141e9b6-a634-4eb5-8e14-79b2526956f5", "raw_diff": "@changed_PLACEMARKER_OPEN Sold Out@changed_PLACEMARKER_CLOSED\n@changed_into_PLACEMARKER_OPENIn Stock: True - Price: 99.0@changed_into_PLACEMARKER_CLOSED", "markup_text_links_to_html_links": false, "restock": {"in_stock": true, "price": 99.0, "currency": "USD", "last_price": 99.0, "availability": "http://schema.org/InStock", "previous_price": null}, "notification_timestamp": 1784178832.2136314, "prev_snapshot": " Sold Out", "preview_url": "https://changedetection.benfou.fr/preview/c141e9b6-a634-4eb5-8e14-79b2526956f5", "screenshot": null, "timestamp_from": null, "timestamp_to": null, "triggered_text": "", "llm_summary": null, "llm_intent": null, "uuid": "c141e9b6-a634-4eb5-8e14-79b2526956f5", "watch_mime_type": "text/html; charset=utf-8", "watch_tag": "Technologie", "watch_title": "AquaPurr 4 - Battery powered cat water fountain - Never needs cleaning", "watch_url": "https://aquapurr.com/products/aquapurr-5-refurbished", "watch_uuid": "c141e9b6-a634-4eb5-8e14-79b2526956f5", "notification_urls": ["jsons://n8n.benfou.fr/webhook/c8bb38ef-b9d8-45d0-837d-c239411db01b/?title={{ watch_title|urlencode }}"], "notification_title": "Changement d\u00e9tect\u00e9 - {{watch_title}}", "notification_body": "{{watch_url}} a chang\u00e9e.\r\n---\r\n{{diff}}\r\n---\r\n", "notification_format": "htmlcolor", "_llm_result": null, "_llm_intent": "", "_llm_change_summary": "", "edit_url": "https://changedetection.benfou.fr/edit/c141e9b6-a634-4eb5-8e14-79b2526956f5"}}]
Thu Jul 16 07:13:34 2026 - SENDING - [{"title": "Changement d\u00e9tect\u00e9 - Update OpenClaw", "body": "https://github.com/openclaw/openclaw/releases a chang\u00e9e. \n--- \n      * openclaw 2026.7.1 \n      * openclaw 2026.7.1-beta.6 \n      * OpenClaw 2026.7.1-beta.5 \n      * openclaw 2026.7.1-beta.2 \n      * openclaw 2026.7.2-beta.1 \n---", "url": "jsons://n8n.benfou.fr/webhook/c8bb38ef-b9d8-45d0-837d-c239411db01b/?title=Update%20OpenClaw&avatar_url=https://raw.githubusercontent.com/dgtlmoon/changedetection.io/master/changedetectionio/static/images/avatar-256x256.png&format=html", "original_context": {"base_url": "https://changedetection.benfou.fr/", "change_datetime": "2026-07-16 07:12:58 CEST", "current_snapshot": " * openclaw 2026.7.2-beta.1", "diff": "@changed_PLACEMARKER_OPEN * openclaw 2026.7.1@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * openclaw 2026.7.1-beta.6@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * OpenClaw 2026.7.1-beta.5@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * openclaw 2026.7.1-beta.2@changed_PLACEMARKER_CLOSED\n@changed_into_PLACEMARKER_OPEN * openclaw 2026.7.2-beta.1@changed_into_PLACEMARKER_CLOSED", "diff_clean": "", "diff_added": "", "diff_added_clean": "", "diff_full": "", "diff_full_clean": "", "diff_patch": "", "diff_removed": "", "diff_removed_clean": "", "diff_changed_from": "", "diff_changed_to": "", "diff_url": "https://changedetection.benfou.fr/diff/27f8a5ca-cd09-4e82-8b3c-5eb340fe0bfc", "raw_diff": "@changed_PLACEMARKER_OPEN * openclaw 2026.7.1@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * openclaw 2026.7.1-beta.6@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * OpenClaw 2026.7.1-beta.5@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * openclaw 2026.7.1-beta.2@changed_PLACEMARKER_CLOSED\n@changed_into_PLACEMARKER_OPEN * openclaw 2026.7.2-beta.1@changed_into_PLACEMARKER_CLOSED", "markup_text_links_to_html_links": false, "restock": {}, "notification_timestamp": 1784178814.2623577, "prev_snapshot": " * openclaw 2026.7.1\n * openclaw 2026.7.1-beta.6\n * OpenClaw 2026.7.1-beta.5\n * openclaw 2026.7.1-beta.2", "preview_url": "https://changedetection.benfou.fr/preview/27f8a5ca-cd09-4e82-8b3c-5eb340fe0bfc", "screenshot": null, "timestamp_from": null, "timestamp_to": null, "triggered_text": "", "llm_summary": null, "llm_intent": null, "uuid": "27f8a5ca-cd09-4e82-8b3c-5eb340fe0bfc", "watch_mime_type": "text/html; charset=utf-8", "watch_tag": "Technologie", "watch_title": "Update OpenClaw", "watch_url": "https://github.com/openclaw/openclaw/releases", "watch_uuid": "27f8a5ca-cd09-4e82-8b3c-5eb340fe0bfc", "notification_urls": ["jsons://n8n.benfou.fr/webhook/c8bb38ef-b9d8-45d0-837d-c239411db01b/?title={{ watch_title|urlencode }}"], "notification_title": "Changement d\u00e9tect\u00e9 - {{watch_title}}", "notification_body": "{{watch_url}} a chang\u00e9e.\r\n---\r\n{{diff}}\r\n---\r\n", "notification_format": "htmlcolor", "_llm_result": null, "_llm_intent": "", "_llm_change_summary": "", "edit_url": "https://changedetection.benfou.fr/edit/27f8a5ca-cd09-4e82-8b3c-5eb340fe0bfc"}}]
Thu Jul 16 07:12:56 2026 - SENDING - [{"title": "Changement d\u00e9tect\u00e9 - API Claude", "body": "https://platform.claude.com/docs/en/about-claude/pricing a chang\u00e9e. \n--- \n    Cookie settings \n \n    We use cookies to deliver and improve our services, analyze site usage, and if you agree, to customize or personalize your experience and market our services to you. You can read our Cookie Policy here. \n \n      Customize Customize Cookie Settings Reject Reject All Cookies Accept Accept All Cookies \n        Claude Platform Docs \n    * Messages \n    * Managed Agents \n    * Admin \n    * Resources \ue027 \n      \ue04a API reference \n      \ue053 English \ue027 \ue04f ConsoleLog in \n      \ue0d3 \ue062 \n            \ue0d3 \n            Search... \n            \u2318K \n                Models \n                Models overview Model IDs and versioning Choosing a model Introducing Claude Fable 5 and Claude Mythos 5 What's new in Claude Opus 4.8 What's new in Claude Sonnet 5 Upgrade between model versions Model deprecations Model cards System prompts Pricing \n            \ue104 \n              Log in \n          \ue0dd Models & pricing \ue027 Pricing \n              Models & pricing / Models \n \n              Pricing \n \n      Cookie settings \n \n      We use cookies to deliver and improve our services, analyze site usage, and if you agree, to customize or personalize your experience and market our services to you. You can read our Cookie Policy here. \n \n        Customize Customize Cookie Settings Reject Reject All Cookies Accept Accept All Cookies \n          Claude Platform Docs \n      * Messages \n      * Managed Agents \n      * Admin \n      * Resources \ue027 \n        \ue04a API reference \n        \ue053 English \ue027 \ue04f ConsoleLog in \n        \ue0d3 \ue062 \n              \ue0d3 \n              Search... \n              \u2318K \n                  Models \n                  Models overview Model IDs and versioning Choosing a model Introducing Claude Fable 5 and Claude Mythos 5 What's new in Claude Opus 4.8 What's new in Claude Sonnet 5 Upgrade between model versions Model deprecations Model cards System prompts Pricing \n              \ue104 \n                Log in \n            \ue0dd Models & pricing \ue027 Pricing \n                Models & pricing / Models \n \n                Pricing \n \n                      Copy page \n                    \ue027 \n                Learn about Anthropic's pricing structure for models and features \n              Learn about Anthropic's pricing structure for models and features \n                  Copy page \n                \ue027 \n \n            This page provides detailed pricing information for Anthropic's models and features. All prices are in USD. \n \n            For the most current pricing information, visit claude.com/pricing. \n \n                \ue09a \n                Model pricing \n \n            The following table shows pricing for all Claude models: \n \n                Model                                                           Base Input Tokens  5m Cache Writes  1h Cache Writes  Cache Hits & Refreshes  Output Tokens \n                Claude Fable 5                                                  $10 / MTok         $12.50 / MTok    $20 / MTok       $1 / MTok               $50 / MTok \n                Claude Mythos 5 (limited availability)                          $10 / MTok         $12.50 / MTok    $20 / MTok       $1 / MTok               $50 / MTok \n                Claude Opus 4.8                                                 $5 / MTok          $6.25 / MTok     $10 / MTok       $0.50 / MTok            $25 / MTok \n                Claude Opus 4.7                                                 $5 / MTok          $6.25 / MTok     $10 / MTok       $0.50 / MTok            $25 / MTok \n                Claude Opus 4.6                                                 $5 / MTok          $6.25 / MTok     $10 / MTok       $0.50 / MTok            $25 / MTok \n                Claude Opus 4.5                                                 $5 / MTok          $6.25 / MTok     $10 / MTok       $0.50 / MTok            $25 / MTok \n                Claude Opus 4.1 (deprecated)                                    $15 / MTok         $18.75 / MTok    $30 / MTok       $1.50 / MTok            $75 / MTok \n                Claude Opus 4 (retired, except on Google Cloud)                 $15 / MTok         $18.75 / MTok    $30 / MTok       $1.50 / MTok            $75 / MTok \n                Claude Sonnet 5                                                 $2 / MTok          $2.50 / MTok     $4 / MTok        $0.20 / MTok            $10 / MTok \n                through August 31, 2026 \n                Claude Sonnet 5                                                 $3 / MTok          $3.75 / MTok     $6 / MTok        $0.30 / MTok            $15 / MTok \n                starting September 1, 2026 \n                Claude Sonnet 4.6                                               $3 / MTok          $3.75 / MTok     $6 / MTok        $0.30 / MTok            $15 / MTok \n                Claude Sonnet 4.5                                               $3 / MTok          $3.75 / MTok     $6 / MTok        $0.30 / MTok            $15 / MTok \n                Claude Sonnet 4 (retired, except on Bedrock and Google Cloud)   $3 / MTok          $3.75 / MTok     $6 / MTok        $0.30 / MTok            $15 / MTok \n                Claude Haiku 4.5                                                $1 / MTok          $1.25 / MTok     $2 / MTok        $0.10 / MTok            $5 / MTok \n                Claude Haiku 3.5 (retired, except on Bedrock and Google Cloud)  $0.80 / MTok       $1 / MTok        $1.60 / MTok     $0.08 / MTok            $4 / MTok \n \n                  \ue08f \n \n                  Introductory pricing of $2/$10 per million input/output tokens is in effect through August 31, 2026, after which the standard pricing of $3/$15 per million input/output tokens will take effect. \n \n                  \ue08f \n \n                  MTok = Million tokens. The "Base Input Tokens" column shows standard input pricing, the "5m Cache Writes", "1h Cache Writes", and "Cache Hits & Refreshes" columns are specific to prompt caching, and "Output Tokens" shows output pricing. See prompt caching pricing for an explanation of the cache columns and pricing multipliers. \n \n                  \ue08f \n \n                  Claude Opus 4.7 and later Opus models, Claude Fable 5, Claude Mythos 5, Claude Mythos Preview, and Claude Sonnet 5 use a newer tokenizer that contributes to their improved performance on a wide range of tasks. This tokenizer produces approximately 30% more tokens for the same text. The exact increase depends on the content and workload shape. Claude Sonnet 4.6 and earlier models use the previous tokenizer. \n \n              For Claude Platform on AWS pricing, see Claude Platform on AWS pricing. \n \n              This page provides detailed pricing information for Anthropic's models and features. All prices are in USD. \n \n              For the most current pricing information, visit claude.com/pricing. \n                  Cloud platform pricing \n \n              This section covers partner-operated cloud platforms, where the cloud provider invoices you. For Anthropic-operated cloud platforms billed through a marketplace, see Claude Platform on AWS pricing and Claude in Microsoft Foundry pricing. \n \n              Claude models are available on Amazon Bedrock and Google Cloud. For official pricing, visit: \n \n                * Amazon Bedrock pricing \n                * Google Cloud pricing \n                  \ue08f \n \n                  Regional and multi-region endpoint pricing for Claude 4.5 models and beyond \n \n                  Starting with Claude Sonnet 4.5, Haiku 4.5, and Opus 4.5: \n \n                    * Bedrock offers two endpoint types: global endpoints (dynamic routing for maximum availability) and regional endpoints (guaranteed data routing through specific geographic regions). \n                    * Google Cloud offers three endpoint types: global endpoints, multi-region endpoints (dynamic routing within a geographic area), and regional endpoints. \n \n                  Regional and multi-region endpoints include a 10% premium over global endpoints. The Claude API (first-party) is global by default; for first-party data residency options and pricing, see Data residency pricing. \n \n                  Scope: This pricing structure applies to Claude Sonnet 4.5, Haiku 4.5, Opus 4.5, and all future models. Earlier models (Claude Opus 4.1 (deprecated) and prior releases) retain their existing pricing. \n \n                  For implementation details and code examples: \n \n                    * Amazon Bedrock global vs regional endpoints for Opus 4.7, Haiku 4.5, and later models, or the legacy integration for all other models on Bedrock \n                    * Google Cloud global, multi-region, and regional endpoints \n \n                  \ue09a \n                  Claude Platform on AWS pricing \n \n              Claude Platform on AWS bills through AWS Marketplace using Claude Consumption Units (CCUs). Anthropic rates your token usage in USD at standard per-model, per-feature rates, applies any negotiated discount, converts the result to CCUs at $0.01 per CCU, and reports the CCU quantity to AWS Marketplace hourly. Your AWS bill shows a single CCU line item. \n \n                  Concept          Details \n                  Billing unit     Claude Consumption Unit (CCU) \n                  CCU price        $0.01 per CCU (fixed; discounts apply at token-to-CCU conversion, not to the CCU price) \n                  Conversion       Token usage rated in USD at standard per-model, per-feature rates (same as Claude API pricing), then converted to CCUs at $0.01 per CCU \n                  Billing cadence  Hourly metering to AWS Marketplace; monthly invoices \n                  Payment model    Arrears only (postpaid); no prepaid credits \n                  Discounts        Applied as fewer CCUs metered \n                  Tax              Pre-tax metering; AWS Marketplace handles tax \n                  Cost visibility  Real-time breakdown in the Claude Console (access through the AWS Console); AWS Cost Explorer shows aggregated CCU \n                  Model pricing \n \n              The following table shows pricing for all Claude models: \n \n                  Model                                                           Base Input Tokens  5m Cache Writes  1h Cache Writes  Cache Hits & Refreshes  Output Tokens \n                  Claude Fable 5                                                  $10 / MTok         $12.50 / MTok    $20 / MTok       $1 / MTok               $50 / MTok \n                  Claude Mythos 5 (limited availability)                          $10 / MTok         $12.50 / MTok    $20 / MTok       $1 / MTok               $50 / MTok \n                  Claude Opus 4.8                                                 $5 / MTok          $6.25 / MTok     $10 / MTok       $0.50 / MTok            $25 / MTok \n                  Claude Opus 4.7                                                 $5 / MTok          $6.25 / MTok     $10 / MTok       $0.50 / MTok            $25 / MTok \n                  Claude Opus 4.6                                                 $5 / MTok          $6.25 / MTok     $10 / MTok       $0.50 / MTok            $25 / MTok \n                  Claude Opus 4.5                                                 $5 / MTok          $6.25 / MTok     $10 / MTok       $0.50 / MTok            $25 / MTok \n                  Claude Opus 4.1 (deprecated)                                    $15 / MTok         $18.75 / MTok    $30 / MTok       $1.50 / MTok            $75 / MTok \n                  Claude Opus 4 (retired, except on Google Cloud)                 $15 / MTok         $18.75 / MTok    $30 / MTok       $1.50 / MTok            $75 / MTok \n                  Claude Sonnet 5                                                 $2 / MTok          $2.50 / MTok     $4 / MTok        $0.20 / MTok            $10 / MTok \n                  through August 31, 2026 \n                  Claude Sonnet 5                                                 $3 / MTok          $3.75 / MTok     $6 / MTok        $0.30 / MTok            $15 / MTok \n                  starting September 1, 2026 \n                  Claude Sonnet 4.6                                               $3 / MTok          $3.75 / MTok     $6 / MTok        $0.30 / MTok            $15 / MTok \n                  Claude Sonnet 4.5                                               $3 / MTok          $3.75 / MTok     $6 / MTok        $0.30 / MTok            $15 / MTok \n                  Claude Sonnet 4 (retired, except on Bedrock and Google Cloud)   $3 / MTok          $3.75 / MTok     $6 / MTok        $0.30 / MTok            $15 / MTok \n                  Claude Haiku 4.5                                                $1 / MTok          $1.25 / MTok     $2 / MTok        $0.10 / MTok            $5 / MTok \n                  Claude Haiku 3.5 (retired, except on Bedrock and Google Cloud)  $0.80 / MTok       $1 / MTok        $1.60 / MTok     $0.08 / MTok            $4 / MTok \n                    Claude Consumption Units. If Customer accesses the Services through certain Marketplace Platforms (e.g., Claude Platform on AWS), usage will be invoiced in Claude Consumption Units ("CCU") rather than per MTok. A CCU is a unit of measure used solely for Marketplace Platform invoicing. One hundred (100) CCU represents $1.00 USD of fees owed for the Services, calculated at the applicable prices on claude.com/pricing#api, after application of any discounts. \n                    Introductory pricing of $2/$10 per million input/output tokens is in effect through August 31, 2026, after which the standard pricing of $3/$15 per million input/output tokens will take effect. \n \n                    \ue08f \n \n                    MTok = Million tokens. The "Base Input Tokens" column shows standard input pricing, the "5m Cache Writes", "1h Cache Writes", and "Cache Hits & Refreshes" columns are specific to prompt caching, and "Output Tokens" shows output pricing. See prompt caching pricing for an explanation of the cache columns and pricing multipliers. \n \n                    \ue08f \n \n                    Claude Opus 4.7 and later Opus models, Claude Fable 5, Claude Mythos 5, Claude Mythos Preview, and Claude Sonnet 5 use a newer tokenizer that contributes to their improved performance on a wide range of tasks. This tokenizer produces approximately 30% more tokens for the same text. The exact increase depends on the content and workload shape. Claude Sonnet 4.6 and earlier models use the previous tokenizer. \n \n                For Claude Platform on AWS pricing, see Claude Platform on AWS pricing. \n                    Inference geography \n \n                For Claude Opus 4.6, Claude Sonnet 4.6, and later models, using inference_geo: "us" applies a 1.1x pricing multiplier. inference_geo: "global" (default) uses standard pricing. See Data residency for details. \n                    Cloud platform pricing \n \n                This section covers partner-operated cloud platforms, where the cloud provider invoices you. For Anthropic-operated cloud platforms billed through a marketplace, see Claude Platform on AWS pricing and Claude in Microsoft Foundry pricing. \n \n                Claude models are available on Amazon Bedrock and Google Cloud. For official pricing, visit: \n \n                  * Amazon Bedrock pricing \n                  * Google Cloud pricing \n                    \ue08f \n \n                    Regional and multi-region endpoint pricing for Claude 4.5 models and beyond \n \n                    Starting with Claude Sonnet 4.5, Haiku 4.5, and Opus 4.5: \n \n                      * Bedrock offers two endpoint types: global endpoints (dynamic routing for maximum availability) and regional endpoints (guaranteed data routing through specific geographic regions). \n                      * Google Cloud offers three endpoint types: global endpoints, multi-region endpoints (dynamic routing within a geographic area), and regional endpoints. \n \n                    Regional and multi-region endpoints include a 10% premium over global endpoints. The Claude API (first-party) is global by default; for first-party data residency options and pricing, see Data residency pricing. \n \n                    Scope: This pricing structure applies to Claude Sonnet 4.5, Haiku 4.5, Opus 4.5, and all future models. Earlier models (Claude Opus 4.1 (deprecated) and prior releases) retain their existing pricing. \n \n                    For implementation details and code examples: \n \n                      * Amazon Bedrock global vs regional endpoints for Opus 4.7, Haiku 4.5, and later models, or the legacy integration for all other models on Bedrock \n                      * Google Cloud global, multi-region, and regional endpoints \n                    Private offers \n \n                When you sign up on the AWS Console Claude Platform on AWS service page, the AWS Console looks up any private offer associated with your account and prompts you to accept it in AWS Marketplace. Contact your Anthropic account representative for private offer terms. \n \n                    \ue08f \n \n                    If you have an existing Amazon Bedrock private offer, contact your Anthropic or AWS account representative before getting started with Claude Platform on AWS to ensure your discounts are applied correctly. Discounts cannot be applied retroactively to usage incurred before your private offer is accepted. \n \n                    \ue09a \n                    Claude in Microsoft Foundry pricing \n \n                Claude in Microsoft Foundry bills through the Azure Marketplace using Claude Consumption Units (CCUs). Anthropic rates your token usage in USD at standard per-model, per-feature rates, applies any negotiated discount, converts the result to CCUs at $0.01 per CCU, and reports the CCU quantity to the Azure Marketplace hourly. Your Azure bill shows a single CCU line item. \n                    Claude Platform on AWS pricing \n \n                Claude Platform on AWS bills through AWS Marketplace using Claude Consumption Units (CCUs). Anthropic rates your token usage in USD at standard per-model, per-feature rates, applies any negotiated discount, converts the result to CCUs at $0.01 per CCU, and reports the CCU quantity to AWS Marketplace hourly. Your AWS bill shows a single CCU line item. \n                    Billing cadence  Hourly metering to the AzureAWS Marketplace; monthly invoices \n                    Tax              Pre-tax metering; Azure Marketplace handles tax \n                    Cost visibility  Azure Cost Management shows aggregated CCU \n                    Tax              Pre-tax metering; AWS Marketplace handles tax \n                    Cost visibility  Real-time breakdown in the Claude Console (access through the AWS Console); AWS Cost Explorer shows aggregated CCU \n                      Claude Consumption Units. If Customer accesses the Services through certain Marketplace Platforms (e.g., Claude Platform on AWS, Claude in Microsoft Foundry),AWS), usage will be invoiced in Claude Consumption Units ("CCU") rather than per MTok. A CCU is a unit of measure used solely for Marketplace Platform invoicing. One hundred (100) CCU represents $1.00 USD of fees owed for the Services, calculated at the applicable prices on claude.com/pricing#api, after application of any discounts. \n                  Deployments hosted on Azure canFor useClaude theOpus US Data4.6, Zone Standard deployment type, which keeps inference within theClaude UnitedSonnet States.4.6, Thisand islater equivalentmodels, tousing inference_geo: "us" onapplies thea Claude1.1x APIpricing andmultiplier. appliesinference_geo: the"global" same(default) 1.1xuses pricingstandard multiplier.pricing. See Data residency for details. \n                      Feature-specific pricing \n                      Private offers \n \n                  When you sign up on the AWS Console Claude Platform on AWS service page, the AWS Console looks up any private offer associated with your account and prompts you to accept it in AWS Marketplace. Contact your Anthropic account representative for private offer terms. \n \n                      \ue08f \n \n                      If you have an existing Amazon Bedrock private offer, contact your Anthropic or AWS account representative before getting started with Claude Platform on AWS to ensure your discounts are applied correctly. Discounts cannot be applied retroactively to usage incurred before your private offer is accepted. \n                      Prompt caching \n \n                  Prompt caching reduces costs and latency by reusing previously processed portions of your prompt across API calls. Instead of reprocessing the same large system prompt, document, or conversation history on every request, the API reads from cache at a fraction of the standard input price. \n \n                  There are two ways to enable prompt caching: \n \n                    * Automatic caching: Add a single cache_control field at the top level of your request. The system automatically manages cache breakpoints as conversations grow. This is the recommended starting point for most use cases. \n                    * Explicit cache breakpoints: Place cache_control directly on individual content blocks for fine-grained control over exactly what gets cached. \n \n                  Prompt caching uses the following pricing multipliers relative to base input token rates: \n \n                      Cache operation       Multiplier              Duration \n                      5-minute cache write  1.25x base input price  Cache valid for 5 minutes \n                      1-hour cache write    2x base input price     Cache valid for 1 hour \n                      Cache read (hit)      0.1x base input price   Same duration as the preceding write \n \n \n                    Cache write tokens are charged when content is first stored. Cache read tokens are charged when a subsequent request retrieves the cached content. A cache hit costs 10% of the standard input price, which means caching pays off after just one cache read for the 5-minute duration (1.25x write), or after two cache reads for the 1-hour duration (2x write). \n \n                    These multipliers stack with other pricing modifiers, including the Batch API discount and data residency. \n \n                    For implementation details, supported models, and code examples, see Prompt caching. \n                      Claude in Microsoft Foundry pricing \n \n                  Claude in Microsoft Foundry bills through the Azure Marketplace using Claude Consumption Units (CCUs). Anthropic rates your token usage in USD at standard per-model, per-feature rates, applies any negotiated discount, converts the result to CCUs at $0.01 per CCU, and reports the CCU quantity to the Azure Marketplace hourly. Your Azure bill shows a single CCU line item. \n \n                      Concept          Details \n                      Billing unit     Claude Consumption Unit (CCU) \n                      CCU price        $0.01 per CCU (fixed; discounts apply at token-to-CCU conversion, not to the CCU price) \n                      Conversion       Token usage rated in USD at standard per-model, per-feature rates (same as Claude API pricing), then converted to CCUs at $0.01 per CCU \n                      Billing cadence  Hourly metering to the Azure Marketplace; monthly invoices \n                      Payment model    Arrears only (postpaid); no prepaid credits \n                      Discounts        Applied as fewer CCUs metered \n                      Tax              Pre-tax metering; Azure Marketplace handles tax \n                      Cost visibility  Azure Cost Management shows aggregated CCU \n \n                        \ue08f \n \n                        Claude Consumption Units. If Customer accesses the Services through certain Marketplace Platforms (e.g., Claude Platform on AWS, Claude in Microsoft Foundry), usage will be invoiced in Claude Consumption Units ("CCU") rather than per MTok. A CCU is a unit of measure used solely for Marketplace Platform invoicing. One hundred (100) CCU represents $1.00 USD of fees owed for the Services, calculated at the applicable prices on claude.com/pricing#api, after application of any discounts. \n                        Data residency pricing \n \n                    For Claude Opus 4.6, Claude Sonnet 4.6, and later models, specifying US-only inference through the inference_geo parameter incurs a 1.1x multiplier on all token pricing categories, including input tokens, output tokens, cache writes, and cache reads. Global routing (the default) uses standard pricing. \n \n                    This applies to the Claude API (first-party) and Claude Platform on AWS. On Claude in Microsoft Foundry, the same 1.1x multiplier applies to deployments that use the US Data Zone Standard deployment type (see Inference geography). Partner-operated platforms (Bedrock and Google Cloud) have independent regional pricing. See Bedrock and Google Cloud for details. Earlier models do not support the inference_geo parameter and always use standard pricing; requests that include the parameter on these models return a 400 error. \n \n                    For more information, see Data residency. \n                        Inference geography \n \n                    Deployments hosted on Azure can use the US Data Zone Standard deployment type, which keeps inference within the United States. This is equivalent to inference_geo: "us" on the Claude API and applies the same 1.1x pricing multiplier. See Data residency for details. \n                        Fast mode pricing \n \n                    Fast mode, in research preview, provides significantly faster output for Claude Opus 4.8 and Claude Opus 4.7 at premium pricing. Fast mode pricing applies across the full context window, including requests over 200k input tokens. Fast mode is not available on Claude Platform on AWS. \n \n                        Model            Input       Output \n                        Claude Opus 4.8  $10 / MTok  $50 / MTok \n                        Claude Opus 4.7  $30 / MTok  $150 / MTok \n \n \n                      Fast mode for Claude Opus 4.7 is deprecated and will be removed on July 24, 2026. As of June 29, 2026, fast mode is not available on Claude Opus 4.6: requests to claude-opus-4-6 with speed: "fast" run at standard speed and are billed at standard rates. See Fast mode. \n \n                      Fast mode pricing stacks with other pricing modifiers: \n \n                        * Prompt caching multipliers apply on top of fast mode pricing \n                        * Data residency multipliers apply on top of fast mode pricing \n \n                      Fast mode is not available with the Batch API. \n \n                      For more information, see Fast mode. \n                        Feature-specific pricing \n \n                        \ue09a \n                        Prompt caching \n \n                    Prompt caching reduces costs and latency by reusing previously processed portions of your prompt across API calls. Instead of reprocessing the same large system prompt, document, or conversation history on every request, the API reads from cache at a fraction of the standard input price. \n \n                    There are two ways to enable prompt caching: \n \n                      * Automatic caching: Add a single cache_control field at the top level of your request. The system automatically manages cache breakpoints as conversations grow. This is the recommended starting point for most use cases. \n                      * Explicit cache breakpoints: Place cache_control directly on individual content blocks for fine-grained control over exactly what gets cached. \n \n                    Prompt caching uses the following pricing multipliers relative to base input token rates: \n \n                        Cache operation       Multiplier              Duration \n                        5-minute cache write  1.25x base input price  Cache valid for 5 minutes \n                        1-hour cache write    2x base input price     Cache valid for 1 hour \n                        Cache read (hit)      0.1x base input price   Same duration as the preceding write \n \n \n                      Cache write tokens are charged when content is first stored. Cache read tokens are charged when a subsequent request retrieves the cached content. A cache hit costs 10% of the standard input price, which means caching pays off after just one cache read for the 5-minute duration (1.25x write), or after two cache reads for the 1-hour duration (2x write). \n \n                      These multipliers stack with other pricing modifiers, including the Batch API discount and data residency. \n \n                      For implementation details, supported models, and code examples, see Prompt caching. \n                          Batch processing \n \n                      The Batch API allows asynchronous processing of large volumes of requests with a 50% discount on both input and output tokens. \n \n                          Model                                                           Batch input   Batch output \n                          Claude Fable 5                                                  $5 / MTok     $25 / MTok \n                          Claude Mythos 5 (limited availability)                          $5 / MTok     $25 / MTok \n                          Claude Opus 4.8                                                 $2.50 / MTok  $12.50 / MTok \n                          Claude Opus 4.7                                                 $2.50 / MTok  $12.50 / MTok \n                          Claude Opus 4.6                                                 $2.50 / MTok  $12.50 / MTok \n                          Claude Opus 4.5                                                 $2.50 / MTok  $12.50 / MTok \n                          Claude Opus 4.1 (deprecated)                                    $7.50 / MTok  $37.50 / MTok \n                          Claude Opus 4 (retired, except on Google Cloud)                 $7.50 / MTok  $37.50 / MTok \n                          Claude Sonnet 5                                                 $1 / MTok     $5 / MTok \n                          through August 31, 2026 \n                          Claude Sonnet 5                                                 $1.50 / MTok  $7.50 / MTok \n                          starting September 1, 2026 \n                          Claude Sonnet 4.6                                               $1.50 / MTok  $7.50 / MTok \n                          Claude Sonnet 4.5                                               $1.50 / MTok  $7.50 / MTok \n                          Claude Sonnet 4 (retired, except on Bedrock and Google Cloud)   $1.50 / MTok  $7.50 / MTok \n                          Claude Haiku 4.5                                                $0.50 / MTok  $2.50 / MTok \n                          Claude Haiku 3.5 (retired, except on Bedrock and Google Cloud)  $0.40 / MTok  $2 / MTok \n \n \n                        For more information about batch processing, see Batch processing. \n                          Data residency pricing \n \n                      For Claude Opus 4.6, Claude Sonnet 4.6, and later models, specifying US-only inference through the inference_geo parameter incurs a 1.1x multiplier on all token pricing categories, including input tokens, output tokens, cache writes, and cache reads. Global routing (the default) uses standard pricing. \n \n                      This applies to the Claude API (first-party) and Claude Platform on AWS. On Claude in Microsoft Foundry, the same 1.1x multiplier applies to deployments that use the US Data Zone Standard deployment type (see Inference geography). Partner-operated platforms (Bedrock and Google Cloud) have independent regional pricing. See Bedrock and Google Cloud for details. Earlier models do not support the inference_geo parameter and always use standard pricing; requests that include the parameter on these models return a 400 error. \n \n                      For more information, see Data residency. \n \n                          \ue09a \n                          Fast mode pricing \n \n                      Fast mode, in research preview, provides significantly faster output for Claude Opus 4.8 and Claude Opus 4.7 at premium pricing. Fast mode pricing applies across the full context window, including requests over 200k input tokens. Fast mode is not available on Claude Platform on AWS. \n \n                          Model            Input       Output \n                          Claude Opus 4.8  $10 / MTok  $50 / MTok \n                          Claude Opus 4.7  $30 / MTok  $150 / MTok \n \n \n                        Fast mode for Claude Opus 4.7 is deprecated and will be removed on July 24, 2026. As of June 29, 2026, fast mode is not available on Claude Opus 4.6: requests to claude-opus-4-6 with speed: "fast" run at standard speed and are billed at standard rates. See Fast mode. \n \n                        Fast mode pricing stacks with other pricing modifiers: \n \n                          * Prompt caching multipliers apply on top of fast mode pricing \n                          * Data residency multipliers apply on top of fast mode pricing \n \n                        Fast mode is not available with the Batch API. \n \n                        For more information, see Fast mode. \n                            Long context pricing \n \n                        Claude Fable 5, Claude Mythos 5, Claude Mythos Preview, Claude Opus 4.8, Opus 4.7, Opus 4.6, Sonnet 5, and Sonnet 4.6 include the full 1M token context window at standard pricing. (A 900k-token request is billed at the same per-token rate as a 9k-token request.) Prompt caching and batch processing discounts apply at standard rates across the full context window. \n \n                            \ue09a \n                            Tool use pricing \n \n                        Tool use requests are priced based on: \n \n                         1. The total number of input tokens sent to the model (including in the tools parameter) \n                         2. The number of output tokens generated \n                         3. For server-side tools, additional usage-based pricing (e.g., web search charges per search performed) \n \n                        Client-side tools are priced the same as any other Claude API request, while server-side tools may incur additional charges based on their specific usage. \n \n                        The additional tokens from tool use come from: \n \n                          * The tools parameter in API requests (tool names, descriptions, and schemas) \n                          * tool_use content blocks in API requests and responses \n                          * tool_result content blocks in API requests \n \n                        When you use tools, the API also automatically includes a special system prompt for the model which enables tool use. The number of tool use tokens required for each model are listed below (excluding the additional tokens listed above). Note that the table assumes at least 1 tool is provided. If no tools are provided, then a tool choice of none uses 0 additional system prompt tokens. \n \n                            Model                                                           Tool choice          Tool use system prompt token count \n                            Claude Opus 4.8                                                 auto, noneany, tool  290 tokens410 tokens \n                            Claude Opus 4.7                                                 auto, noneany, tool  675 tokens804 tokens \n                            Claude Opus 4.6                                                 auto, noneany, tool  497 tokens589 tokens \n                            Claude Opus 4.5                                                 auto, noneany, tool  496 tokens588 tokens \n                            Claude Opus 4.1 (deprecated)                                    auto, noneany, tool  313 tokens315 tokens \n                            Claude Opus 4 (retired, except on Google Cloud)                 auto, noneany, tool  313 tokens315 tokens \n                            Claude Sonnet 5                                                 auto, noneany, tool  354 tokens474 tokens \n                            Claude Sonnet 4.6                                               auto, noneany, tool  497 tokens589 tokens \n                            Claude Sonnet 4.5                                               auto, noneany, tool  496 tokens588 tokens \n                            Claude Sonnet 4 (retired, except on Bedrock and Google Cloud)   auto, noneany, tool  313 tokens315 tokens \n                            Claude Haiku 4.5                                                auto, noneany, tool  496 tokens588 tokens \n                            Claude Haiku 3.5 (retired, except on Bedrock and Google Cloud)  auto, noneany, tool  264 tokens355 tokens \n \n \n                          These token counts are added to your normal input and output tokens to calculate the total cost of a request. \n \n                          For current per-model prices, refer to the model pricing section. \n \n                          For more information about tool use implementation and best practices, see Tool use. \n                            Batch processing \n \n                        The Batch API allows asynchronous processing of large volumes of requests with a 50% discount on both input and output tokens. \n \n                            Model                                                           Batch input   Batch output \n                            Claude Fable 5                                                  $5 / MTok     $25 / MTok \n                            Claude Mythos 5 (limited availability)                          $5 / MTok     $25 / MTok \n                            Claude Opus 4.8                                                 $2.50 / MTok  $12.50 / MTok \n                            Claude Opus 4.7                                                 $2.50 / MTok  $12.50 / MTok \n                            Claude Opus 4.6                                                 $2.50 / MTok  $12.50 / MTok \n                            Claude Opus 4.5                                                 $2.50 / MTok  $12.50 / MTok \n                            Claude Opus 4.1 (deprecated)                                    $7.50 / MTok  $37.50 / MTok \n                            Claude Opus 4 (retired, except on Google Cloud)                 $7.50 / MTok  $37.50 / MTok \n                            Claude Sonnet 5                                                 $1 / MTok     $5 / MTok \n                            through August 31, 2026 \n                            Claude Sonnet 5                                                 $1.50 / MTok  $7.50 / MTok \n                            starting September 1, 2026 \n                            Claude Sonnet 4.6                                               $1.50 / MTok  $7.50 / MTok \n                            Claude Sonnet 4.5                                               $1.50 / MTok  $7.50 / MTok \n                            Claude Sonnet 4 (retired, except on Bedrock and Google Cloud)   $1.50 / MTok  $7.50 / MTok \n                            Claude Haiku 4.5                                                $0.50 / MTok  $2.50 / MTok \n                            Claude Haiku 3.5 (retired, except on Bedrock and Google Cloud)  $0.40 / MTok  $2 / MTok \n \n \n                          For more information about batch processing, see Batch processing. \n                              Specific tool pricing \n                              Long context pricing \n \n                          Claude Fable 5, Claude Mythos 5, Claude Mythos Preview, Claude Opus 4.8, Opus 4.7, Opus 4.6, Sonnet 5, and Sonnet 4.6 include the full 1M token context window at standard pricing. (A 900k-token request is billed at the same per-token rate as a 9k-token request.) Prompt caching and batch processing discounts apply at standard rates across the full context window. \n                              Bash tool \n \n                          The bash tool adds 245 input tokens to your API calls. \n \n                          Additional tokens are consumed by: \n \n                            * Command outputs (stdout/stderr) \n                            * Error messages \n                            * Large file contents \n \n                          See tool use pricing for complete pricing details. \n \n                              \ue09a \n                              Code execution tool \n \n                          Code execution is free when used with web search or web fetch. When web_search_20260209 (or later) or web_fetch_20260209 (or later) is included in your API request, there are no additional charges for code execution tool calls beyond the standard input and output token costs. \n \n                          When used without these tools, code execution is billed by execution time, tracked separately from token usage: \n \n                            * Execution time has a minimum of 5 minutes \n                            * Each organization receives 1,550 free hours of usage per month \n                            * Additional usage beyond 1,550 hours is billed at $0.05 per hour, per container \n                            * If files are included in the request, execution time is billed even if the tool is not invoked, due to files being preloaded onto the container \n \n                          Code execution usage is tracked in the response: \n \n                                { "usage" : { "input_tokens" : 105 , "output_tokens" : 239 , "server_tool_use" : { "code_execution_requests" : 1 } } } \n                              \ue056 \n \n                              \ue09a \n                              Text editor tool \n \n                          The text editor tool uses the same pricing structure as other tools used with Claude. It follows the standard input and output token pricing based on the Claude model you're using. \n \n                          In addition to the base tokens, the following additional input tokens are needed for the text editor tool: \n \n                              Tool                               Additional input tokens \n                              text_editor_20250429 (Claude 4.x)  700 tokens \n \n \n                            See tool use pricing for complete pricing details. \n                              Tool use pricing \n \n                          Tool use requests are priced based on: \n \n                           1. The total number of input tokens sent to the model (including in the tools parameter) \n                           2. The number of output tokens generated \n                           3. For server-side tools, additional usage-based pricing (e.g., web search charges per search performed) \n \n                          Client-side tools are priced the same as any other Claude API request, while server-side tools may incur additional charges based on their specific usage. \n \n                          The additional tokens from tool use come from: \n \n                            * The tools parameter in API requests (tool names, descriptions, and schemas) \n                            * tool_use content blocks in API requests and responses \n                            * tool_result content blocks in API requests \n \n                          When you use tools, the API also automatically includes a special system prompt for the model which enables tool use. The number of tool use tokens required for each model are listed below (excluding the additional tokens listed above). Note that the table assumes at least 1 tool is provided. If no tools are provided, then a tool choice of none uses 0 additional system prompt tokens. \n \n                              Model                                                           Tool choice          Tool use system prompt token count \n                              Claude Opus 4.8                                                 auto, noneany, tool  290 tokens410 tokens \n                              Claude Opus 4.7                                                 auto, noneany, tool  675 tokens804 tokens \n                              Claude Opus 4.6                                                 auto, noneany, tool  497 tokens589 tokens \n                              Claude Opus 4.5                                                 auto, noneany, tool  496 tokens588 tokens \n                              Claude Opus 4.1 (deprecated)                                    auto, noneany, tool  313 tokens315 tokens \n                              Claude Opus 4 (retired, except on Google Cloud)                 auto, noneany, tool  313 tokens315 tokens \n                              Claude Sonnet 5                                                 auto, noneany, tool  354 tokens474 tokens \n                              Claude Sonnet 4.6                                               auto, noneany, tool  497 tokens589 tokens \n                              Claude Sonnet 4.5                                               auto, noneany, tool  496 tokens588 tokens \n                              Claude Sonnet 4 (retired, except on Bedrock and Google Cloud)   auto, noneany, tool  313 tokens315 tokens \n                              Claude Haiku 4.5                                                auto, noneany, tool  496 tokens588 tokens \n                              Claude Haiku 3.5 (retired, except on Bedrock and Google Cloud)  auto, noneany, tool  264 tokens355 tokens \n \n \n                            These token counts are added to your normal input and output tokens to calculate the total cost of a request. \n \n                            For current per-model prices, refer to the model pricing section. \n \n                            For more information about tool use implementation and best practices, see Tool use. \n                                Web search tool \n \n                            Web search usage is charged in addition to token usage: \n \n                                  { "usage" : { "input_tokens" : 105 , "output_tokens" : 6039 , "cache_read_input_tokens" : 7123 , "cache_creation_input_tokens" : 7345 , "server_tool_use" : { "web_search_requests" : 1 } } } \n                                \ue056 \n \n                            Web search is available on the Claude API for $10 per 1,000 searches, plus standard token costs for search-generated content. Web search results retrieved throughout a conversation are counted as input tokens, in search iterations executed during a single turn and in subsequent conversation turns. \n \n                            Each web search counts as one use, regardless of the number of results returned. If an error occurs during web search, the web search will not be billed. \n                                Specific tool pricing \n                                Web fetch tool \n \n                            Web fetch usage has no additional charges beyond standard token costs: \n \n                                  { "usage" : { "input_tokens" : 25039 , "output_tokens" : 931 , "cache_read_input_tokens" : 0 , "cache_creation_input_tokens" : 0 , "server_tool_use" : { "web_fetch_requests" : 1 } } } \n                                \ue056 \n \n                            The web fetch tool is available on the Claude API at no additional cost. You only pay standard token costs for the fetched content that becomes part of your conversation context. \n \n                            To protect against inadvertently fetching large content that would consume excessive tokens, use the max_content_tokens parameter to set appropriate limits based on your use case and budget considerations. \n \n                            Example token usage for typical content: \n \n                              * Average web page (10 kB): ~2,500 tokens \n                              * Large documentation page (100 kB): ~25,000 tokens \n                              * Research paper PDF (500 kB): ~125,000 tokens \n \n                                \ue09a \n                                Computer use tool \n \n                            Computer use follows the standard tool use pricing. When using the computer use tool: \n \n                            System prompt overhead: The computer use beta adds 466-499 tokens to the system prompt \n \n                            Computer use tool token usage: \n \n                                Model              Input tokens per tool definition \n                                Claude 4.x models  735 tokens \n \n \n                              Additional token consumption: \n \n                                * Screenshot images (see Vision pricing) \n                                * Tool execution results returned to Claude \n                                  \ue08f \n \n                                  If you're also using bash or text editor tools alongside computer use, those tools have their own token costs as documented in their respective pages. \n                                Bash tool \n \n                            The bash tool definition adds the following input tokens to your request. This is in addition to the per-model tool use system prompt that applies whenever any tool is present. \n \n                                Model                                            Additional input tokens \n                                Claude Opus 4.7 and Claude Opus 4.8              325 tokens \n                                Claude Opus 4.6, Claude Sonnet 4.6, and earlier  244 tokens \n \n \n                              Additional tokens are consumed by: \n \n                                * Command outputs (stdout/stderr) \n                                * Error messages \n                                * Large file contents \n \n                              See tool use pricing for complete pricing details. \n                                  Claude Managed Agents pricing \n \n                              Claude Managed Agents is billed on two dimensions: tokens and session runtime. \n                                  Code execution tool \n \n                              Code execution is free when used with web search or web fetch. When web_search_20260209 (or later) or web_fetch_20260209 (or later) is included in your API request, there are no additional charges for code execution tool calls beyond the standard input and output token costs. \n \n                              When used without these tools, code execution is billed by execution time, tracked separately from token usage: \n \n                                * Execution time has a minimum of 5 minutes \n                                * Each organization receives 1,550 free hours of usage per month \n                                * Additional usage beyond 1,550 hours is billed at $0.05 per hour, per container \n                                * If files are included in the request, execution time is billed even if the tool is not invoked, due to files being preloaded onto the container \n \n                              Code execution usage is tracked in the response: \n \n                                    { "usage" : { "input_tokens" : 105 , "output_tokens" : 239 , "server_tool_use" : { "code_execution_requests" : 1 } } } \n                                  \ue056 \n                                  Tokens \n \n                              All tokens consumed by a Claude Managed Agents session are billed at the rates shown in Model pricing. Prompt caching multipliers apply identically. Web search triggered inside a session incurs the standard $10 per 1,000 searches. On Claude Platform on AWS, session token and runtime charges convert to Claude Consumption Units at the standard rate. \n \n                              The following Messages API modifiers do not apply to Claude Managed Agents sessions: \n \n                                  Modifier                   Why it doesn't apply \n                                  Batch API discount         Sessions are stateful and interactive. There is no batch mode. \n                                  Fast mode premium          Inference speed is managed by the runtime. \n                                  Data residency multiplier  inference_geo is a Messages API request field. \n                                  Cloud platform pricing     Not available on partner-operated cloud platforms. \n \n                                  Text editor tool \n \n                              The text editor tool uses the same pricing structure as other tools used with Claude. It follows the standard input and output token pricing based on the Claude model you're using. \n \n                              In addition to the base tokens, the following additional input tokens are needed for the text editor tool: \n \n                                  Tool                               Additional input tokens \n                                  text_editor_20250429 (Claude 4.x)  700 tokens \n \n \n                                See tool use pricing for complete pricing details. \n                                    Session runtime \n \n                                    SKU              Rate                    Metering \n                                    Session runtime  $0.08 per session-hour  running status duration \n \n \n                                  Runtime is measured to the millisecond and accrues only while the session's status is running. Time spent idle (waiting for your next message or a tool confirmation), rescheduling, or terminated does not count toward runtime. \n \n                                    Web search tool \n \n                                Web search usage is charged in addition to token usage: \n \n                                      { "usage" : { "input_tokens" : 105 , "output_tokens" : 6039 , "cache_read_input_tokens" : 7123 , "cache_creation_input_tokens" : 7345 , "server_tool_use" : { "web_search_requests" : 1 } } } \n                                    \ue056 \n \n                                Web search is available on the Claude API for $10 per 1,000 searches, plus standard token costs for search-generated content. Web search results retrieved throughout a conversation are counted as input tokens, in search iterations executed during a single turn and in subsequent conversation turns. \n \n                                Each web search counts as one use, regardless of the number of results returned. If an error occurs during web search, the web search will not be billed. \n \n                                    \ue09a \n                                    Web fetch tool \n \n                                Web fetch usage has no additional charges beyond standard token costs: \n \n                                      { "usage" : { "input_tokens" : 25039 , "output_tokens" : 931 , "cache_read_input_tokens" : 0 , "cache_creation_input_tokens" : 0 , "server_tool_use" : { "web_fetch_requests" : 1 } } } \n                                    \ue056 \n \n                                The web fetch tool is available on the Claude API at no additional cost. You only pay standard token costs for the fetched content that becomes part of your conversation context. \n \n                                To protect against inadvertently fetching large content that would consume excessive tokens, use the max_content_tokens parameter to set appropriate limits based on your use case and budget considerations. \n \n                                Example token usage for typical content: \n \n                                  * Average web page (10 kB): ~2,500 tokens \n                                  * Large documentation page (100 kB): ~25,000 tokens \n                                  * Research paper PDF (500 kB): ~125,000 tokens \n \n                                    \ue09a \n                                    Computer use tool \n \n                                Computer use follows the standard tool use pricing. When using the computer use tool: \n \n                                System prompt overhead: The computer use beta adds 466-499 tokens to the system prompt \n \n                                Computer use tool token usage: \n \n                                    Model              Input tokens per tool definition \n                                    Claude 4.x models  735 tokens \n \n \n                                  Additional token consumption: \n \n                                    * Screenshot images (see Vision pricing) \n                                    * Tool execution results returned to Claude \n                                      Session runtime replaces the Code Execution container-hour billing model when using Claude Managed Agents. You are not separately billed for container hours on top of session runtime. \n                                      If you're also using bash or text editor tools alongside computer use, those tools have their own token costs as documented in their respective pages. \n                                      Worked example \n \n                                  A one-hour coding session using Claude Opus 4.8 that consumes 50,000 input tokens and 15,000 output tokens: \n \n                                      Line item        Calculation               Cost \n                                      Input tokens     50,000 \u00d7 $5 / 1,000,000   $0.25 \n                                      Output tokens    15,000 \u00d7 $25 / 1,000,000  $0.375 \n                                      Session runtime  1.0 hour \u00d7 $0.08          $0.08 \n                                      Total                                      $0.705 \n \n \n                                    If prompt caching is active and 40,000 of the input tokens are cache reads: \n \n                                        Line item              Calculation                    Cost \n                                        Uncached input tokens  10,000 \u00d7 $5 / 1,000,000        $0.05 \n                                        Cache read tokens      40,000 \u00d7 $5 \u00d7 0.1 / 1,000,000  $0.02 \n                                        Output tokens          15,000 \u00d7 $25 / 1,000,000       $0.375 \n                                        Session runtime        1.0 hour \u00d7 $0.08               $0.08 \n                                        Total                                                 $0.525 \n                                      Claude Managed Agents pricing \n \n                                  Claude Managed Agents is billed on two dimensions: tokens and session runtime. \n \n                                      \ue09a \n                                      Tokens \n \n                                  All tokens consumed by a Claude Managed Agents session are billed at the rates shown in Model pricing. Prompt caching multipliers apply identically. Web search triggered inside a session incurs the standard $10 per 1,000 searches. On Claude Platform on AWS, session token and runtime charges convert to Claude Consumption Units at the standard rate. \n \n                                  The following Messages API modifiers do not apply to Claude Managed Agents sessions: \n \n                                      Modifier                   Why it doesn't apply \n                                      Batch API discount         Sessions are stateful and interactive. There is no batch mode. \n                                      Fast mode premium          Inference speed is managed by the runtime. \n                                      Data residency multiplier  inference_geo is a Messages API request field. \n                                      Cloud platform pricing     Not available on partner-operated cloud platforms. \n \n \n                                        \ue09a \n                                        Session runtime \n \n                                        SKU              Rate                    Metering \n                                        Session runtime  $0.08 per session-hour  running status duration \n \n \n                                      Runtime is measured to the millisecond and accrues only while the session's status is running. Time spent idle (waiting for your next message or a tool confirmation), rescheduling, or terminated does not count toward runtime. \n                                          Example calculation for processing 10,000 support tickets: \n \n                                            * Average ~3,700 tokens per conversation \n                                            * Using Claude Haiku 4.5 at $1/MTok input, $5/MTok output \n                                            * Total cost: ~$37.00 per 10,000 tickets \n \n                                      For a detailed walkthrough of this calculation, see the customer support agent guide. \n                                          Session runtime replaces the Code Execution container-hour billing model when using Claude Managed Agents. You are not separately billed for container hours on top of session runtime. \n                                          Additional pricing considerations \n \n                                          \ue09a \n                                          Cost optimization strategies \n \n                                      When building agents with Claude: \n \n                                       1. Use appropriate models: Choose Haiku for simple tasks, Sonnet for most production workloads, and Opus for the most complex reasoning \n                                       2. Implement prompt caching: Reduce costs for repeated context \n                                       3. Batch operations: Use the Batch API for non-time-sensitive tasks \n                                       4. Monitor usage patterns: Track token consumption to identify optimization opportunities \n                                          \ue097 \n \n                                          For high-volume agent applications, contact the enterprise sales team for custom pricing arrangements. \n \n                                          \ue09a \n                                          Rate limits \n \n                                      Rate limits vary by usage tier and affect how many requests you can make: \n \n                                        * Start tier: Entry-level limits for getting started \n                                        * Build tier: Increased limits for growing applications \n                                        * Scale tier: Highest standard limits for production workloads \n \n                                      For detailed rate limit information, see Rate limits. \n \n                                      For limits beyond the Scale tier or custom pricing arrangements, contact the sales team. \n \n                                          \ue09a \n                                          Volume discounts \n \n                                      Volume discounts may be available for high-volume users. These are negotiated on a case-by-case basis. \n \n                                        * Standard usage tiers use the pricing shown in Model pricing \n                                        * Enterprise customers can contact sales for custom pricing \n                                        * Academic and research discounts may be available \n \n                                          \ue09a \n                                          Enterprise pricing \n \n                                      For enterprise customers with specific needs: \n \n                                        * Custom rate limits \n                                        * Volume discounts \n                                        * Dedicated support \n                                        * Custom terms \n \n                                      Contact the sales team at sales@anthropic.com or through the Claude Console to discuss enterprise pricing options. \n \n                                          \ue09a \n                                          Billing and payment \n \n                                        * Billing is based on actual monthly usage \n                                        * All payments are in USD \n                                        * Credit card and invoicing options available \n                                        * Usage tracking available in the Claude Console \n \n                                          \ue09a \n                                          Frequently asked questions \n \n                                      How is token usage calculated? \n \n                                      Tokens are pieces of text that models process. As a rough estimate, 1 token is approximately 4 characters or 0.75 words in English. The exact count varies by language and content type. \n \n                                      Are there free tiers or trials? \n \n                                      New users receive a small amount of free credits to test the API. Contact sales for information about extended trials for enterprise evaluation. \n \n                                      How do discounts stack? \n \n                                      Batch API and prompt caching discounts can be combined. For example, using both features together provides significant cost savings compared to standard API calls. See prompt caching pricing for how the multipliers interact. \n \n                                      What payment methods are accepted? \n \n                                      Major credit cards are accepted for standard accounts. Enterprise customers can arrange invoicing and other payment methods. \n \n                                      For additional questions about pricing, contact support@anthropic.com. \n \n                                        Was this page helpful? \n \n                                          \ue0fb \ue0f9 \n                                          * Model pricing \n                                          * Cloud platform pricing \n                                          * Claude Platform on AWS pricing \n                                          * Inference geography \n                                          * Private offers \n                                          * Claude in Microsoft Foundry pricing \n                                          * Inference geography \n                                          * Feature-specific pricing \n                                          * Prompt caching \n                                          * Data residency pricing \n                                          * Fast mode pricing \n                                          * Batch processing \n                                          * Long context pricing \n                                          * Tool use pricing \n                                          * Specific tool pricing \n                                          * Claude Managed Agents pricing \n                                          * Tokens \n                                          * Session runtime \n                                          * Worked example \n                                          * Additional pricing considerations \n                                          * Cost optimization strategies \n                                          * Rate limits \n                                          * Volume discounts \n                                          * Enterprise pricing \n                                          * Billing and payment \n                                          * Frequently asked questions \n                                            Claude Platform Docs \n \n                                        Solutions \n \n                                          * AI agents \n                                          * Code modernization \n                                          * Coding \n                                          * Customer support \n                                          * Education \n                                          * Financial services \n                                          * Government \n                                          * Life sciences \n \n                                        Partners \n \n                                          * Claude on AWS \n                                          * Claude on Google Cloud \n \n                                        Learn \n \n                                          * Blog \n                                          * Courses \n                                          * Use cases \n                                          * Connectors \n                                          * Customer stories \n                                          * Engineering at Anthropic \n                                          * Events \n                                          * Powered by Claude \n                                          * Service partners \n                                          * Startups program \n \n                                        Company \n \n                                          * Anthropic \n                                          * Careers \n                                          * Economic Futures \n                                          * Research \n                                          * News \n                                          * Responsible Scaling Policy \n                                          * Security and compliance \n                                          * Transparency \n \n                                      Learn \n \n                                        * Blog \n                                        * Courses \n                                        * Use cases \n                                        * Connectors \n                                        * Customer stories \n                                        * Engineering at Anthropic \n                                        * Events \n                                        * Powered by Claude \n                                        * Service partners \n                                        * Startups program \n \n                                        Help and security \n \n                                          * Availability \n                                          * Status \n                                          * Support \n                                          * Discord \n \n                                        Terms and policies \n \n                                          * Privacy policy \n                                          * Responsible disclosure policy \n                                          * Terms of service: Commercial \n                                          * Terms of service: Consumer \n                                          * Usage policy \n                                          Worked example \n \n                                      A one-hour coding session using Claude Opus 4.8 that consumes 50,000 input tokens and 15,000 output tokens: \n \n                                          Line item        Calculation               Cost \n                                          Input tokens     50,000 \u00d7 $5 / 1,000,000   $0.25 \n                                          Output tokens    15,000 \u00d7 $25 / 1,000,000  $0.375 \n                                          Session runtime  1.0 hour \u00d7 $0.08          $0.08 \n                                          Total                                      $0.705 \n \n \n                                        If prompt caching is active and 40,000 of the input tokens are cache reads: \n \n                                            Line item              Calculation                    Cost \n                                            Uncached input tokens  10,000 \u00d7 $5 / 1,000,000        $0.05 \n                                            Cache read tokens      40,000 \u00d7 $5 \u00d7 0.1 / 1,000,000  $0.02 \n                                            Output tokens          15,000 \u00d7 $25 / 1,000,000       $0.375 \n                                            Session runtime        1.0 hour \u00d7 $0.08               $0.08 \n                                            Total                                                 $0.525 \n \n                                              \ue08f \n \n                                              Example calculation for processing 10,000 support tickets: \n \n                                                * Average ~3,700 tokens per conversation \n                                                * Using Claude Haiku 4.5 at $1/MTok input, $5/MTok output \n                                                * Total cost: ~$37.00 per 10,000 tickets \n \n                                          For a detailed walkthrough of this calculation, see the customer support agent guide. \n \n                                              \ue09a \n                                              Additional pricing considerations \n \n                                              \ue09a \n                                              Cost optimization strategies \n \n                                          When building agents with Claude: \n \n                                           1. Use appropriate models: Choose Haiku for simple tasks, Sonnet for most production workloads, and Opus for the most complex reasoning \n                                           2. Implement prompt caching: Reduce costs for repeated context \n                                           3. Batch operations: Use the Batch API for non-time-sensitive tasks \n                                           4. Monitor usage patterns: Track token consumption to identify optimization opportunities \n                                              \ue097 \n \n                                              For high-volume agent applications, contact the enterprise sales team for custom pricing arrangements. \n \n                                              \ue09a \n                                              Rate limits \n \n                                          Rate limits vary by usage tier and affect how many requests you can make: \n \n                                            * Start tier: Entry-level limits for getting started \n                                            * Build tier: Increased limits for growing applications \n                                            * Scale tier: Highest standard limits for production workloads \n \n                                          For detailed rate limit information, see Rate limits. \n \n                                          For limits beyond the Scale tier or custom pricing arrangements, contact the sales team. \n \n                                              \ue09a \n                                              Volume discounts \n \n                                          Volume discounts may be available for high-volume users. These are negotiated on a case-by-case basis. \n \n                                            * Standard usage tiers use the pricing shown in Model pricing \n                                            * Enterprise customers can contact sales for custom pricing \n                                            * Academic and research discounts may be available \n \n                                              \ue09a \n                                              Enterprise pricing \n \n                                          For enterprise customers with specific needs: \n \n                                            * Custom rate limits \n                                            * Volume discounts \n                                            * Dedicated support \n                                            * Custom terms \n \n                                          Contact the sales team at sales@anthropic.com or through the Claude Console to discuss enterprise pricing options. \n \n                                              \ue09a \n                                              Billing and payment \n \n                                            * Billing is based on actual monthly usage \n                                            * All payments are in USD \n                                            * Credit card and invoicing options available \n                                            * Usage tracking available in the Claude Console \n \n                                              \ue09a \n                                              Frequently asked questions \n \n                                          How is token usage calculated? \n \n                                          Tokens are pieces of text that models process. As a rough estimate, 1 token is approximately 4 characters or 0.75 words in English. The exact count varies by language and content type. \n \n                                          Are there free tiers or trials? \n \n                                          New users receive a small amount of free credits to test the API. Contact sales for information about extended trials for enterprise evaluation. \n \n                                          How do discounts stack? \n \n                                          Batch API and prompt caching discounts can be combined. For example, using both features together provides significant cost savings compared to standard API calls. See prompt caching pricing for how the multipliers interact. \n \n                                          What payment methods are accepted? \n \n                                          Major credit cards are accepted for standard accounts. Enterprise customers can arrange invoicing and other payment methods. \n \n                                          For additional questions about pricing, contact support@anthropic.com. \n \n                                            Was this page helpful? \n \n                                              \ue0fb \ue0f9 \n                                              * Model pricing \n                                              * Cloud platform pricing \n                                              * Claude Platform on AWS pricing \n                                              * Inference geography \n                                              * Private offers \n                                              * Claude in Microsoft Foundry pricing \n                                              * Inference geography \n                                              * Feature-specific pricing \n                                              * Prompt caching \n                                              * Data residency pricing \n                                              * Fast mode pricing \n                                              * Batch processing \n                                              * Long context pricing \n                                              * Tool use pricing \n                                              * Specific tool pricing \n                                              * Claude Managed Agents pricing \n                                              * Tokens \n                                              * Session runtime \n                                              * Worked example \n                                              * Additional pricing considerations \n                                              * Cost optimization strategies \n                                              * Rate limits \n                                              * Volume discounts \n                                              * Enterprise pricing \n                                              * Billing and payment \n                                              * Frequently asked questions \n                                                Claude Platform Docs \n \n                                            Solutions \n \n                                              * AI agents \n                                              * Code modernization \n                                              * Coding \n                                              * Customer support \n                                              * Education \n                                              * Financial services \n                                              * Government \n                                              * Life sciences \n \n                                            Partners \n \n                                              * Claude on AWS \n                                              * Claude on Google Cloud \n \n                                            Learn \n \n                                              * Blog \n                                              * Courses \n                                              * Use cases \n                                              * Connectors \n                                              * Customer stories \n                                              * Engineering at Anthropic \n                                              * Events \n                                              * Powered by Claude \n                                              * Service partners \n                                              * Startups program \n \n                                            Company \n \n                                              * Anthropic \n                                              * Careers \n                                              * Economic Futures \n                                              * Research \n                                              * News \n                                              * Responsible Scaling Policy \n                                              * Security and compliance \n                                              * Transparency \n \n                                          Learn \n \n                                            * Blog \n                                            * Courses \n                                            * Use cases \n                                            * Connectors \n                                            * Customer stories \n                                            * Engineering at Anthropic \n                                            * Events \n                                            * Powered by Claude \n                                            * Service partners \n                                            * Startups program \n \n                                            Help and security \n \n                                              * Availability \n                                              * Status \n                                              * Support \n                                              * Discord \n \n                                            Terms and policies \n \n                                              * Privacy policy \n                                              * Responsible disclosure policy \n                                              * Terms of service: Commercial \n                                              * Terms of service: Consumer \n                                              * Usage policy \n                                * If files are included in the request, execution time is billed even if the tool is not invoked, due tobecause files beingare preloaded onto the container \n---", "url": "jsons://n8n.benfou.fr/webhook/c8bb38ef-b9d8-45d0-837d-c239411db01b/?title=API%20Claude&avatar_url=https://raw.githubusercontent.com/dgtlmoon/changedetection.io/master/changedetectionio/static/images/avatar-256x256.png&format=html", "original_context": {"base_url": "https://changedetection.benfou.fr/", "change_datetime": "2026-07-16 07:12:16 CEST", "current_snapshot": " * If files are included in the request, execution time is billed even if the tool is not invoked, due tobecause files beingare preloaded onto the container", "diff": "@changed_PLACEMARKER_OPEN Cookie settings@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN We use cookies to deliver and improve our services, analyze site usage, and if you agree, to customize or personalize your experience and market our services to you. You can read our Cookie Policy here.@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Customize Customize Cookie Settings Reject Reject All Cookies Accept Accept All Cookies@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Claude Platform Docs@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Messages@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Managed Agents@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Admin@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Resources \ue027@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN \ue04a API reference@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN \ue053 English \ue027 \ue04f ConsoleLog in@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN \ue0d3 \ue062@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN \ue0d3@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Search...@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN \u2318K@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Models@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Models overview Model IDs and versioning Choosing a model Introducing Claude Fable 5 and Claude Mythos 5 What's new in Claude Opus 4.8 What's new in Claude Sonnet 5 Upgrade between model versions Model deprecations Model cards System prompts Pricing@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN \ue104@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Log in@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN \ue0dd Models & pricing \ue027 Pricing@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Models & pricing / Models@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Pricing@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Cookie settings@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN We use cookies to deliver and improve our services, analyze site usage, and if you agree, to customize or personalize your experience and market our services to you. You can read our Cookie Policy here.@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Customize Customize Cookie Settings Reject Reject All Cookies Accept Accept All Cookies@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Claude Platform Docs@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Messages@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Managed Agents@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Admin@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Resources \ue027@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN \ue04a API reference@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN \ue053 English \ue027 \ue04f ConsoleLog in@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN \ue0d3 \ue062@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN \ue0d3@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Search...@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN \u2318K@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Models@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Models overview Model IDs and versioning Choosing a model Introducing Claude Fable 5 and Claude Mythos 5 What's new in Claude Opus 4.8 What's new in Claude Sonnet 5 Upgrade between model versions Model deprecations Model cards System prompts Pricing@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN \ue104@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Log in@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN \ue0dd Models & pricing \ue027 Pricing@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Models & pricing / Models@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Pricing@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Copy page@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN \ue027@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Learn about Anthropic's pricing structure for models and features@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Learn about Anthropic's pricing structure for models and features@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Copy page@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN \ue027@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN This page provides detailed pricing information for Anthropic's models and features. All prices are in USD.@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN For the most current pricing information, visit claude.com/pricing.@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN \ue09a@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Model pricing@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN The following table shows pricing for all Claude models:@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Model Base Input Tokens 5m Cache Writes 1h Cache Writes Cache Hits & Refreshes Output Tokens@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Claude Fable 5 $10 / MTok $12.50 / MTok $20 / MTok $1 / MTok $50 / MTok@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Claude Mythos 5 (limited availability) $10 / MTok $12.50 / MTok $20 / MTok $1 / MTok $50 / MTok@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Claude Opus 4.8 $5 / MTok $6.25 / MTok $10 / MTok $0.50 / MTok $25 / MTok@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Claude Opus 4.7 $5 / MTok $6.25 / MTok $10 / MTok $0.50 / MTok $25 / MTok@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Claude Opus 4.6 $5 / MTok $6.25 / MTok $10 / MTok $0.50 / MTok $25 / MTok@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Claude Opus 4.5 $5 / MTok $6.25 / MTok $10 / MTok $0.50 / MTok $25 / MTok@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Claude Opus 4.1 (deprecated) $15 / MTok $18.75 / MTok $30 / MTok $1.50 / MTok $75 / MTok@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Claude Opus 4 (retired, except on Google Cloud) $15 / MTok $18.75 / MTok $30 / MTok $1.50 / MTok $75 / MTok@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Claude Sonnet 5 $2 / MTok $2.50 / MTok $4 / MTok $0.20 / MTok $10 / MTok@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN through August 31, 2026@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Claude Sonnet 5 $3 / MTok $3.75 / MTok $6 / MTok $0.30 / MTok $15 / MTok@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN starting September 1, 2026@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Claude Sonnet 4.6 $3 / MTok $3.75 / MTok $6 / MTok $0.30 / MTok $15 / MTok@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Claude Sonnet 4.5 $3 / MTok $3.75 / MTok $6 / MTok $0.30 / MTok $15 / MTok@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Claude Sonnet 4 (retired, except on Bedrock and Google Cloud) $3 / MTok $3.75 / MTok $6 / MTok $0.30 / MTok $15 / MTok@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Claude Haiku 4.5 $1 / MTok $1.25 / MTok $2 / MTok $0.10 / MTok $5 / MTok@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Claude Haiku 3.5 (retired, except on Bedrock and Google Cloud) $0.80 / MTok $1 / MTok $1.60 / MTok $0.08 / MTok $4 / MTok@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN \ue08f@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Introductory pricing of $2/$10 per million input/output tokens is in effect through August 31, 2026, after which the standard pricing of $3/$15 per million input/output tokens will take effect.@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN \ue08f@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN MTok = Million tokens. The "Base Input Tokens" column shows standard input pricing, the "5m Cache Writes", "1h Cache Writes", and "Cache Hits & Refreshes" columns are specific to prompt caching, and "Output Tokens" shows output pricing. See prompt caching pricing for an explanation of the cache columns and pricing multipliers.@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN \ue08f@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Claude Opus 4.7 and later Opus models, Claude Fable 5, Claude Mythos 5, Claude Mythos Preview, and Claude Sonnet 5 use a newer tokenizer that contributes to their improved performance on a wide range of tasks. This tokenizer produces approximately 30% more tokens for the same text. The exact increase depends on the content and workload shape. Claude Sonnet 4.6 and earlier models use the previous tokenizer.@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN For Claude Platform on AWS pricing, see Claude Platform on AWS pricing.@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN This page provides detailed pricing information for Anthropic's models and features. All prices are in USD.@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN For the most current pricing information, visit claude.com/pricing.@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Cloud platform pricing@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN This section covers partner-operated cloud platforms, where the cloud provider invoices you. For Anthropic-operated cloud platforms billed through a marketplace, see Claude Platform on AWS pricing and Claude in Microsoft Foundry pricing.@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Claude models are available on Amazon Bedrock and Google Cloud. For official pricing, visit:@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Amazon Bedrock pricing@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Google Cloud pricing@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN \ue08f@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Regional and multi-region endpoint pricing for Claude 4.5 models and beyond@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Starting with Claude Sonnet 4.5, Haiku 4.5, and Opus 4.5:@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Bedrock offers two endpoint types: global endpoints (dynamic routing for maximum availability) and regional endpoints (guaranteed data routing through specific geographic regions).@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Google Cloud offers three endpoint types: global endpoints, multi-region endpoints (dynamic routing within a geographic area), and regional endpoints.@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Regional and multi-region endpoints include a 10% premium over global endpoints. The Claude API (first-party) is global by default; for first-party data residency options and pricing, see Data residency pricing.@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Scope: This pricing structure applies to Claude Sonnet 4.5, Haiku 4.5, Opus 4.5, and all future models. Earlier models (Claude Opus 4.1 (deprecated) and prior releases) retain their existing pricing.@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN For implementation details and code examples:@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Amazon Bedrock global vs regional endpoints for Opus 4.7, Haiku 4.5, and later models, or the legacy integration for all other models on Bedrock@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Google Cloud global, multi-region, and regional endpoints@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN \ue09a@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Claude Platform on AWS pricing@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Claude Platform on AWS bills through AWS Marketplace using Claude Consumption Units (CCUs). Anthropic rates your token usage in USD at standard per-model, per-feature rates, applies any negotiated discount, converts the result to CCUs at $0.01 per CCU, and reports the CCU quantity to AWS Marketplace hourly. Your AWS bill shows a single CCU line item.@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Concept Details@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Billing unit Claude Consumption Unit (CCU)@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN CCU price $0.01 per CCU (fixed; discounts apply at token-to-CCU conversion, not to the CCU price)@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Conversion Token usage rated in USD at standard per-model, per-feature rates (same as Claude API pricing), then converted to CCUs at $0.01 per CCU@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Billing cadence Hourly metering to AWS Marketplace; monthly invoices@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Payment model Arrears only (postpaid); no prepaid credits@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Discounts Applied as fewer CCUs metered@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Tax Pre-tax metering; AWS Marketplace handles tax@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Cost visibility Real-time breakdown in the Claude Console (access through the AWS Console); AWS Cost Explorer shows aggregated CCU@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Model pricing@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN The following table shows pricing for all Claude models:@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Model Base Input Tokens 5m Cache Writes 1h Cache Writes Cache Hits & Refreshes Output Tokens@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Claude Fable 5 $10 / MTok $12.50 / MTok $20 / MTok $1 / MTok $50 / MTok@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Claude Mythos 5 (limited availability) $10 / MTok $12.50 / MTok $20 / MTok $1 / MTok $50 / MTok@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Claude Opus 4.8 $5 / MTok $6.25 / MTok $10 / MTok $0.50 / MTok $25 / MTok@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Claude Opus 4.7 $5 / MTok $6.25 / MTok $10 / MTok $0.50 / MTok $25 / MTok@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Claude Opus 4.6 $5 / MTok $6.25 / MTok $10 / MTok $0.50 / MTok $25 / MTok@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Claude Opus 4.5 $5 / MTok $6.25 / MTok $10 / MTok $0.50 / MTok $25 / MTok@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Claude Opus 4.1 (deprecated) $15 / MTok $18.75 / MTok $30 / MTok $1.50 / MTok $75 / MTok@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Claude Opus 4 (retired, except on Google Cloud) $15 / MTok $18.75 / MTok $30 / MTok $1.50 / MTok $75 / MTok@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Claude Sonnet 5 $2 / MTok $2.50 / MTok $4 / MTok $0.20 / MTok $10 / MTok@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN through August 31, 2026@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Claude Sonnet 5 $3 / MTok $3.75 / MTok $6 / MTok $0.30 / MTok $15 / MTok@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN starting September 1, 2026@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Claude Sonnet 4.6 $3 / MTok $3.75 / MTok $6 / MTok $0.30 / MTok $15 / MTok@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Claude Sonnet 4.5 $3 / MTok $3.75 / MTok $6 / MTok $0.30 / MTok $15 / MTok@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Claude Sonnet 4 (retired, except on Bedrock and Google Cloud) $3 / MTok $3.75 / MTok $6 / MTok $0.30 / MTok $15 / MTok@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Claude Haiku 4.5 $1 / MTok $1.25 / MTok $2 / MTok $0.10 / MTok $5 / MTok@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Claude Haiku 3.5 (retired, except on Bedrock and Google Cloud) $0.80 / MTok $1 / MTok $1.60 / MTok $0.08 / MTok $4 / MTok@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Claude Consumption Units. If Customer accesses the Services through certain Marketplace Platforms (e.g., Claude Platform on AWS), usage will be invoiced in Claude Consumption Units ("CCU") rather than per MTok. A CCU is a unit of measure used solely for Marketplace Platform invoicing. One hundred (100) CCU represents $1.00 USD of fees owed for the Services, calculated at the applicable prices on claude.com/pricing#api, after application of any discounts.@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Introductory pricing of $2/$10 per million input/output tokens is in effect through August 31, 2026, after which the standard pricing of $3/$15 per million input/output tokens will take effect.@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN \ue08f@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN MTok = Million tokens. The "Base Input Tokens" column shows standard input pricing, the "5m Cache Writes", "1h Cache Writes", and "Cache Hits & Refreshes" columns are specific to prompt caching, and "Output Tokens" shows output pricing. See prompt caching pricing for an explanation of the cache columns and pricing multipliers.@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN \ue08f@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Claude Opus 4.7 and later Opus models, Claude Fable 5, Claude Mythos 5, Claude Mythos Preview, and Claude Sonnet 5 use a newer tokenizer that contributes to their improved performance on a wide range of tasks. This tokenizer produces approximately 30% more tokens for the same text. The exact increase depends on the content and workload shape. Claude Sonnet 4.6 and earlier models use the previous tokenizer.@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN For Claude Platform on AWS pricing, see Claude Platform on AWS pricing.@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Inference geography@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN For Claude Opus 4.6, Claude Sonnet 4.6, and later models, using inference_geo: "us" applies a 1.1x pricing multiplier. inference_geo: "global" (default) uses standard pricing. See Data residency for details.@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Cloud platform pricing@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN This section covers partner-operated cloud platforms, where the cloud provider invoices you. For Anthropic-operated cloud platforms billed through a marketplace, see Claude Platform on AWS pricing and Claude in Microsoft Foundry pricing.@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Claude models are available on Amazon Bedrock and Google Cloud. For official pricing, visit:@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Amazon Bedrock pricing@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Google Cloud pricing@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN \ue08f@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Regional and multi-region endpoint pricing for Claude 4.5 models and beyond@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Starting with Claude Sonnet 4.5, Haiku 4.5, and Opus 4.5:@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Bedrock offers two endpoint types: global endpoints (dynamic routing for maximum availability) and regional endpoints (guaranteed data routing through specific geographic regions).@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Google Cloud offers three endpoint types: global endpoints, multi-region endpoints (dynamic routing within a geographic area), and regional endpoints.@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Regional and multi-region endpoints include a 10% premium over global endpoints. The Claude API (first-party) is global by default; for first-party data residency options and pricing, see Data residency pricing.@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Scope: This pricing structure applies to Claude Sonnet 4.5, Haiku 4.5, Opus 4.5, and all future models. Earlier models (Claude Opus 4.1 (deprecated) and prior releases) retain their existing pricing.@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN For implementation details and code examples:@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Amazon Bedrock global vs regional endpoints for Opus 4.7, Haiku 4.5, and later models, or the legacy integration for all other models on Bedrock@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Google Cloud global, multi-region, and regional endpoints@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Private offers@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN When you sign up on the AWS Console Claude Platform on AWS service page, the AWS Console looks up any private offer associated with your account and prompts you to accept it in AWS Marketplace. Contact your Anthropic account representative for private offer terms.@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN \ue08f@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN If you have an existing Amazon Bedrock private offer, contact your Anthropic or AWS account representative before getting started with Claude Platform on AWS to ensure your discounts are applied correctly. Discounts cannot be applied retroactively to usage incurred before your private offer is accepted.@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN \ue09a@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Claude in Microsoft Foundry pricing@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Claude in Microsoft Foundry bills through the Azure Marketplace using Claude Consumption Units (CCUs). Anthropic rates your token usage in USD at standard per-model, per-feature rates, applies any negotiated discount, converts the result to CCUs at $0.01 per CCU, and reports the CCU quantity to the Azure Marketplace hourly. Your Azure bill shows a single CCU line item.@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Claude Platform on AWS pricing@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Claude Platform on AWS bills through AWS Marketplace using Claude Consumption Units (CCUs). Anthropic rates your token usage in USD at standard per-model, per-feature rates, applies any negotiated discount, converts the result to CCUs at $0.01 per CCU, and reports the CCU quantity to AWS Marketplace hourly. Your AWS bill shows a single CCU line item.@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Billing cadence Hourly metering to the AzureAWS Marketplace; monthly invoices@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Tax Pre-tax metering; Azure Marketplace handles tax@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Cost visibility Azure Cost Management shows aggregated CCU@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Tax Pre-tax metering; AWS Marketplace handles tax@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Cost visibility Real-time breakdown in the Claude Console (access through the AWS Console); AWS Cost Explorer shows aggregated CCU@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Claude Consumption Units. If Customer accesses the Services through certain Marketplace Platforms (e.g., Claude Platform on AWS, Claude in Microsoft Foundry),AWS), usage will be invoiced in Claude Consumption Units ("CCU") rather than per MTok. A CCU is a unit of measure used solely for Marketplace Platform invoicing. One hundred (100) CCU represents $1.00 USD of fees owed for the Services, calculated at the applicable prices on claude.com/pricing#api, after application of any discounts.@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Deployments hosted on Azure canFor useClaude theOpus US Data4.6, Zone Standard deployment type, which keeps inference within theClaude UnitedSonnet States.4.6, Thisand islater equivalentmodels, tousing inference_geo: "us" onapplies thea Claude1.1x APIpricing andmultiplier. appliesinference_geo: the"global" same(default) 1.1xuses pricingstandard multiplier.pricing. See Data residency for details.@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Feature-specific pricing@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Private offers@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN When you sign up on the AWS Console Claude Platform on AWS service page, the AWS Console looks up any private offer associated with your account and prompts you to accept it in AWS Marketplace. Contact your Anthropic account representative for private offer terms.@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN \ue08f@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN If you have an existing Amazon Bedrock private offer, contact your Anthropic or AWS account representative before getting started with Claude Platform on AWS to ensure your discounts are applied correctly. Discounts cannot be applied retroactively to usage incurred before your private offer is accepted.@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Prompt caching@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Prompt caching reduces costs and latency by reusing previously processed portions of your prompt across API calls. Instead of reprocessing the same large system prompt, document, or conversation history on every request, the API reads from cache at a fraction of the standard input price.@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN There are two ways to enable prompt caching:@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Automatic caching: Add a single cache_control field at the top level of your request. The system automatically manages cache breakpoints as conversations grow. This is the recommended starting point for most use cases.@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Explicit cache breakpoints: Place cache_control directly on individual content blocks for fine-grained control over exactly what gets cached.@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Prompt caching uses the following pricing multipliers relative to base input token rates:@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Cache operation Multiplier Duration@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN 5-minute cache write 1.25x base input price Cache valid for 5 minutes@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN 1-hour cache write 2x base input price Cache valid for 1 hour@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Cache read (hit) 0.1x base input price Same duration as the preceding write@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Cache write tokens are charged when content is first stored. Cache read tokens are charged when a subsequent request retrieves the cached content. A cache hit costs 10% of the standard input price, which means caching pays off after just one cache read for the 5-minute duration (1.25x write), or after two cache reads for the 1-hour duration (2x write).@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN These multipliers stack with other pricing modifiers, including the Batch API discount and data residency.@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN For implementation details, supported models, and code examples, see Prompt caching.@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Claude in Microsoft Foundry pricing@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Claude in Microsoft Foundry bills through the Azure Marketplace using Claude Consumption Units (CCUs). Anthropic rates your token usage in USD at standard per-model, per-feature rates, applies any negotiated discount, converts the result to CCUs at $0.01 per CCU, and reports the CCU quantity to the Azure Marketplace hourly. Your Azure bill shows a single CCU line item.@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Concept Details@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Billing unit Claude Consumption Unit (CCU)@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN CCU price $0.01 per CCU (fixed; discounts apply at token-to-CCU conversion, not to the CCU price)@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Conversion Token usage rated in USD at standard per-model, per-feature rates (same as Claude API pricing), then converted to CCUs at $0.01 per CCU@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Billing cadence Hourly metering to the Azure Marketplace; monthly invoices@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Payment model Arrears only (postpaid); no prepaid credits@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Discounts Applied as fewer CCUs metered@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Tax Pre-tax metering; Azure Marketplace handles tax@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Cost visibility Azure Cost Management shows aggregated CCU@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN \ue08f@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Claude Consumption Units. If Customer accesses the Services through certain Marketplace Platforms (e.g., Claude Platform on AWS, Claude in Microsoft Foundry), usage will be invoiced in Claude Consumption Units ("CCU") rather than per MTok. A CCU is a unit of measure used solely for Marketplace Platform invoicing. One hundred (100) CCU represents $1.00 USD of fees owed for the Services, calculated at the applicable prices on claude.com/pricing#api, after application of any discounts.@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Data residency pricing@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN For Claude Opus 4.6, Claude Sonnet 4.6, and later models, specifying US-only inference through the inference_geo parameter incurs a 1.1x multiplier on all token pricing categories, including input tokens, output tokens, cache writes, and cache reads. Global routing (the default) uses standard pricing.@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN This applies to the Claude API (first-party) and Claude Platform on AWS. On Claude in Microsoft Foundry, the same 1.1x multiplier applies to deployments that use the US Data Zone Standard deployment type (see Inference geography). Partner-operated platforms (Bedrock and Google Cloud) have independent regional pricing. See Bedrock and Google Cloud for details. Earlier models do not support the inference_geo parameter and always use standard pricing; requests that include the parameter on these models return a 400 error.@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN For more information, see Data residency.@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Inference geography@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Deployments hosted on Azure can use the US Data Zone Standard deployment type, which keeps inference within the United States. This is equivalent to inference_geo: "us" on the Claude API and applies the same 1.1x pricing multiplier. See Data residency for details.@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Fast mode pricing@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Fast mode, in research preview, provides significantly faster output for Claude Opus 4.8 and Claude Opus 4.7 at premium pricing. Fast mode pricing applies across the full context window, including requests over 200k input tokens. Fast mode is not available on Claude Platform on AWS.@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Model Input Output@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Claude Opus 4.8 $10 / MTok $50 / MTok@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Claude Opus 4.7 $30 / MTok $150 / MTok@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Fast mode for Claude Opus 4.7 is deprecated and will be removed on July 24, 2026. As of June 29, 2026, fast mode is not available on Claude Opus 4.6: requests to claude-opus-4-6 with speed: "fast" run at standard speed and are billed at standard rates. See Fast mode.@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Fast mode pricing stacks with other pricing modifiers:@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Prompt caching multipliers apply on top of fast mode pricing@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Data residency multipliers apply on top of fast mode pricing@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Fast mode is not available with the Batch API.@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN For more information, see Fast mode.@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Feature-specific pricing@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN \ue09a@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Prompt caching@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Prompt caching reduces costs and latency by reusing previously processed portions of your prompt across API calls. Instead of reprocessing the same large system prompt, document, or conversation history on every request, the API reads from cache at a fraction of the standard input price.@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN There are two ways to enable prompt caching:@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Automatic caching: Add a single cache_control field at the top level of your request. The system automatically manages cache breakpoints as conversations grow. This is the recommended starting point for most use cases.@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Explicit cache breakpoints: Place cache_control directly on individual content blocks for fine-grained control over exactly what gets cached.@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Prompt caching uses the following pricing multipliers relative to base input token rates:@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Cache operation Multiplier Duration@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN 5-minute cache write 1.25x base input price Cache valid for 5 minutes@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN 1-hour cache write 2x base input price Cache valid for 1 hour@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Cache read (hit) 0.1x base input price Same duration as the preceding write@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Cache write tokens are charged when content is first stored. Cache read tokens are charged when a subsequent request retrieves the cached content. A cache hit costs 10% of the standard input price, which means caching pays off after just one cache read for the 5-minute duration (1.25x write), or after two cache reads for the 1-hour duration (2x write).@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN These multipliers stack with other pricing modifiers, including the Batch API discount and data residency.@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN For implementation details, supported models, and code examples, see Prompt caching.@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Batch processing@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN The Batch API allows asynchronous processing of large volumes of requests with a 50% discount on both input and output tokens.@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Model Batch input Batch output@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Claude Fable 5 $5 / MTok $25 / MTok@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Claude Mythos 5 (limited availability) $5 / MTok $25 / MTok@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Claude Opus 4.8 $2.50 / MTok $12.50 / MTok@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Claude Opus 4.7 $2.50 / MTok $12.50 / MTok@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Claude Opus 4.6 $2.50 / MTok $12.50 / MTok@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Claude Opus 4.5 $2.50 / MTok $12.50 / MTok@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Claude Opus 4.1 (deprecated) $7.50 / MTok $37.50 / MTok@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Claude Opus 4 (retired, except on Google Cloud) $7.50 / MTok $37.50 / MTok@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Claude Sonnet 5 $1 / MTok $5 / MTok@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN through August 31, 2026@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Claude Sonnet 5 $1.50 / MTok $7.50 / MTok@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN starting September 1, 2026@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Claude Sonnet 4.6 $1.50 / MTok $7.50 / MTok@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Claude Sonnet 4.5 $1.50 / MTok $7.50 / MTok@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Claude Sonnet 4 (retired, except on Bedrock and Google Cloud) $1.50 / MTok $7.50 / MTok@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Claude Haiku 4.5 $0.50 / MTok $2.50 / MTok@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Claude Haiku 3.5 (retired, except on Bedrock and Google Cloud) $0.40 / MTok $2 / MTok@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN For more information about batch processing, see Batch processing.@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Data residency pricing@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN For Claude Opus 4.6, Claude Sonnet 4.6, and later models, specifying US-only inference through the inference_geo parameter incurs a 1.1x multiplier on all token pricing categories, including input tokens, output tokens, cache writes, and cache reads. Global routing (the default) uses standard pricing.@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN This applies to the Claude API (first-party) and Claude Platform on AWS. On Claude in Microsoft Foundry, the same 1.1x multiplier applies to deployments that use the US Data Zone Standard deployment type (see Inference geography). Partner-operated platforms (Bedrock and Google Cloud) have independent regional pricing. See Bedrock and Google Cloud for details. Earlier models do not support the inference_geo parameter and always use standard pricing; requests that include the parameter on these models return a 400 error.@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN For more information, see Data residency.@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN \ue09a@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Fast mode pricing@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Fast mode, in research preview, provides significantly faster output for Claude Opus 4.8 and Claude Opus 4.7 at premium pricing. Fast mode pricing applies across the full context window, including requests over 200k input tokens. Fast mode is not available on Claude Platform on AWS.@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Model Input Output@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Claude Opus 4.8 $10 / MTok $50 / MTok@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Claude Opus 4.7 $30 / MTok $150 / MTok@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Fast mode for Claude Opus 4.7 is deprecated and will be removed on July 24, 2026. As of June 29, 2026, fast mode is not available on Claude Opus 4.6: requests to claude-opus-4-6 with speed: "fast" run at standard speed and are billed at standard rates. See Fast mode.@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Fast mode pricing stacks with other pricing modifiers:@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Prompt caching multipliers apply on top of fast mode pricing@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Data residency multipliers apply on top of fast mode pricing@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Fast mode is not available with the Batch API.@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN For more information, see Fast mode.@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Long context pricing@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Claude Fable 5, Claude Mythos 5, Claude Mythos Preview, Claude Opus 4.8, Opus 4.7, Opus 4.6, Sonnet 5, and Sonnet 4.6 include the full 1M token context window at standard pricing. (A 900k-token request is billed at the same per-token rate as a 9k-token request.) Prompt caching and batch processing discounts apply at standard rates across the full context window.@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN \ue09a@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Tool use pricing@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Tool use requests are priced based on:@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN 1. The total number of input tokens sent to the model (including in the tools parameter)@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN 2. The number of output tokens generated@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN 3. For server-side tools, additional usage-based pricing (e.g., web search charges per search performed)@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Client-side tools are priced the same as any other Claude API request, while server-side tools may incur additional charges based on their specific usage.@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN The additional tokens from tool use come from:@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * The tools parameter in API requests (tool names, descriptions, and schemas)@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * tool_use content blocks in API requests and responses@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * tool_result content blocks in API requests@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN When you use tools, the API also automatically includes a special system prompt for the model which enables tool use. The number of tool use tokens required for each model are listed below (excluding the additional tokens listed above). Note that the table assumes at least 1 tool is provided. If no tools are provided, then a tool choice of none uses 0 additional system prompt tokens.@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Model Tool choice Tool use system prompt token count@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Claude Opus 4.8 auto, noneany, tool 290 tokens410 tokens@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Claude Opus 4.7 auto, noneany, tool 675 tokens804 tokens@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Claude Opus 4.6 auto, noneany, tool 497 tokens589 tokens@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Claude Opus 4.5 auto, noneany, tool 496 tokens588 tokens@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Claude Opus 4.1 (deprecated) auto, noneany, tool 313 tokens315 tokens@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Claude Opus 4 (retired, except on Google Cloud) auto, noneany, tool 313 tokens315 tokens@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Claude Sonnet 5 auto, noneany, tool 354 tokens474 tokens@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Claude Sonnet 4.6 auto, noneany, tool 497 tokens589 tokens@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Claude Sonnet 4.5 auto, noneany, tool 496 tokens588 tokens@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Claude Sonnet 4 (retired, except on Bedrock and Google Cloud) auto, noneany, tool 313 tokens315 tokens@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Claude Haiku 4.5 auto, noneany, tool 496 tokens588 tokens@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Claude Haiku 3.5 (retired, except on Bedrock and Google Cloud) auto, noneany, tool 264 tokens355 tokens@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN These token counts are added to your normal input and output tokens to calculate the total cost of a request.@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN For current per-model prices, refer to the model pricing section.@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN For more information about tool use implementation and best practices, see Tool use.@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Batch processing@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN The Batch API allows asynchronous processing of large volumes of requests with a 50% discount on both input and output tokens.@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Model Batch input Batch output@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Claude Fable 5 $5 / MTok $25 / MTok@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Claude Mythos 5 (limited availability) $5 / MTok $25 / MTok@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Claude Opus 4.8 $2.50 / MTok $12.50 / MTok@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Claude Opus 4.7 $2.50 / MTok $12.50 / MTok@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Claude Opus 4.6 $2.50 / MTok $12.50 / MTok@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Claude Opus 4.5 $2.50 / MTok $12.50 / MTok@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Claude Opus 4.1 (deprecated) $7.50 / MTok $37.50 / MTok@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Claude Opus 4 (retired, except on Google Cloud) $7.50 / MTok $37.50 / MTok@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Claude Sonnet 5 $1 / MTok $5 / MTok@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN through August 31, 2026@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Claude Sonnet 5 $1.50 / MTok $7.50 / MTok@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN starting September 1, 2026@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Claude Sonnet 4.6 $1.50 / MTok $7.50 / MTok@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Claude Sonnet 4.5 $1.50 / MTok $7.50 / MTok@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Claude Sonnet 4 (retired, except on Bedrock and Google Cloud) $1.50 / MTok $7.50 / MTok@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Claude Haiku 4.5 $0.50 / MTok $2.50 / MTok@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Claude Haiku 3.5 (retired, except on Bedrock and Google Cloud) $0.40 / MTok $2 / MTok@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN For more information about batch processing, see Batch processing.@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Specific tool pricing@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Long context pricing@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Claude Fable 5, Claude Mythos 5, Claude Mythos Preview, Claude Opus 4.8, Opus 4.7, Opus 4.6, Sonnet 5, and Sonnet 4.6 include the full 1M token context window at standard pricing. (A 900k-token request is billed at the same per-token rate as a 9k-token request.) Prompt caching and batch processing discounts apply at standard rates across the full context window.@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Bash tool@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN The bash tool adds 245 input tokens to your API calls.@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Additional tokens are consumed by:@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Command outputs (stdout/stderr)@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Error messages@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Large file contents@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN See tool use pricing for complete pricing details.@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN \ue09a@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Code execution tool@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Code execution is free when used with web search or web fetch. When web_search_20260209 (or later) or web_fetch_20260209 (or later) is included in your API request, there are no additional charges for code execution tool calls beyond the standard input and output token costs.@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN When used without these tools, code execution is billed by execution time, tracked separately from token usage:@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Execution time has a minimum of 5 minutes@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Each organization receives 1,550 free hours of usage per month@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Additional usage beyond 1,550 hours is billed at $0.05 per hour, per container@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * If files are included in the request, execution time is billed even if the tool is not invoked, due to files being preloaded onto the container@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Code execution usage is tracked in the response:@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN { "usage" : { "input_tokens" : 105 , "output_tokens" : 239 , "server_tool_use" : { "code_execution_requests" : 1 } } }@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN \ue056@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN \ue09a@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Text editor tool@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN The text editor tool uses the same pricing structure as other tools used with Claude. It follows the standard input and output token pricing based on the Claude model you're using.@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN In addition to the base tokens, the following additional input tokens are needed for the text editor tool:@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Tool Additional input tokens@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN text_editor_20250429 (Claude 4.x) 700 tokens@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN See tool use pricing for complete pricing details.@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Tool use pricing@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Tool use requests are priced based on:@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN 1. The total number of input tokens sent to the model (including in the tools parameter)@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN 2. The number of output tokens generated@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN 3. For server-side tools, additional usage-based pricing (e.g., web search charges per search performed)@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Client-side tools are priced the same as any other Claude API request, while server-side tools may incur additional charges based on their specific usage.@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN The additional tokens from tool use come from:@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * The tools parameter in API requests (tool names, descriptions, and schemas)@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * tool_use content blocks in API requests and responses@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * tool_result content blocks in API requests@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN When you use tools, the API also automatically includes a special system prompt for the model which enables tool use. The number of tool use tokens required for each model are listed below (excluding the additional tokens listed above). Note that the table assumes at least 1 tool is provided. If no tools are provided, then a tool choice of none uses 0 additional system prompt tokens.@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Model Tool choice Tool use system prompt token count@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Claude Opus 4.8 auto, noneany, tool 290 tokens410 tokens@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Claude Opus 4.7 auto, noneany, tool 675 tokens804 tokens@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Claude Opus 4.6 auto, noneany, tool 497 tokens589 tokens@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Claude Opus 4.5 auto, noneany, tool 496 tokens588 tokens@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Claude Opus 4.1 (deprecated) auto, noneany, tool 313 tokens315 tokens@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Claude Opus 4 (retired, except on Google Cloud) auto, noneany, tool 313 tokens315 tokens@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Claude Sonnet 5 auto, noneany, tool 354 tokens474 tokens@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Claude Sonnet 4.6 auto, noneany, tool 497 tokens589 tokens@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Claude Sonnet 4.5 auto, noneany, tool 496 tokens588 tokens@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Claude Sonnet 4 (retired, except on Bedrock and Google Cloud) auto, noneany, tool 313 tokens315 tokens@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Claude Haiku 4.5 auto, noneany, tool 496 tokens588 tokens@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Claude Haiku 3.5 (retired, except on Bedrock and Google Cloud) auto, noneany, tool 264 tokens355 tokens@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN These token counts are added to your normal input and output tokens to calculate the total cost of a request.@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN For current per-model prices, refer to the model pricing section.@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN For more information about tool use implementation and best practices, see Tool use.@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Web search tool@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Web search usage is charged in addition to token usage:@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN { "usage" : { "input_tokens" : 105 , "output_tokens" : 6039 , "cache_read_input_tokens" : 7123 , "cache_creation_input_tokens" : 7345 , "server_tool_use" : { "web_search_requests" : 1 } } }@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN \ue056@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Web search is available on the Claude API for $10 per 1,000 searches, plus standard token costs for search-generated content. Web search results retrieved throughout a conversation are counted as input tokens, in search iterations executed during a single turn and in subsequent conversation turns.@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Each web search counts as one use, regardless of the number of results returned. If an error occurs during web search, the web search will not be billed.@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Specific tool pricing@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Web fetch tool@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Web fetch usage has no additional charges beyond standard token costs:@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN { "usage" : { "input_tokens" : 25039 , "output_tokens" : 931 , "cache_read_input_tokens" : 0 , "cache_creation_input_tokens" : 0 , "server_tool_use" : { "web_fetch_requests" : 1 } } }@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN \ue056@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN The web fetch tool is available on the Claude API at no additional cost. You only pay standard token costs for the fetched content that becomes part of your conversation context.@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN To protect against inadvertently fetching large content that would consume excessive tokens, use the max_content_tokens parameter to set appropriate limits based on your use case and budget considerations.@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Example token usage for typical content:@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Average web page (10 kB): ~2,500 tokens@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Large documentation page (100 kB): ~25,000 tokens@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Research paper PDF (500 kB): ~125,000 tokens@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN \ue09a@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Computer use tool@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Computer use follows the standard tool use pricing. When using the computer use tool:@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN System prompt overhead: The computer use beta adds 466-499 tokens to the system prompt@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Computer use tool token usage:@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Model Input tokens per tool definition@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Claude 4.x models 735 tokens@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Additional token consumption:@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Screenshot images (see Vision pricing)@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Tool execution results returned to Claude@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN \ue08f@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN If you're also using bash or text editor tools alongside computer use, those tools have their own token costs as documented in their respective pages.@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Bash tool@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN The bash tool definition adds the following input tokens to your request. This is in addition to the per-model tool use system prompt that applies whenever any tool is present.@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Model Additional input tokens@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Claude Opus 4.7 and Claude Opus 4.8 325 tokens@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Claude Opus 4.6, Claude Sonnet 4.6, and earlier 244 tokens@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Additional tokens are consumed by:@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Command outputs (stdout/stderr)@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Error messages@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Large file contents@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN See tool use pricing for complete pricing details.@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Claude Managed Agents pricing@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Claude Managed Agents is billed on two dimensions: tokens and session runtime.@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Code execution tool@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Code execution is free when used with web search or web fetch. When web_search_20260209 (or later) or web_fetch_20260209 (or later) is included in your API request, there are no additional charges for code execution tool calls beyond the standard input and output token costs.@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN When used without these tools, code execution is billed by execution time, tracked separately from token usage:@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Execution time has a minimum of 5 minutes@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Each organization receives 1,550 free hours of usage per month@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Additional usage beyond 1,550 hours is billed at $0.05 per hour, per container@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * If files are included in the request, execution time is billed even if the tool is not invoked, due to files being preloaded onto the container@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Code execution usage is tracked in the response:@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN { "usage" : { "input_tokens" : 105 , "output_tokens" : 239 , "server_tool_use" : { "code_execution_requests" : 1 } } }@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN \ue056@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Tokens@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN All tokens consumed by a Claude Managed Agents session are billed at the rates shown in Model pricing. Prompt caching multipliers apply identically. Web search triggered inside a session incurs the standard $10 per 1,000 searches. On Claude Platform on AWS, session token and runtime charges convert to Claude Consumption Units at the standard rate.@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN The following Messages API modifiers do not apply to Claude Managed Agents sessions:@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Modifier Why it doesn't apply@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Batch API discount Sessions are stateful and interactive. There is no batch mode.@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Fast mode premium Inference speed is managed by the runtime.@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Data residency multiplier inference_geo is a Messages API request field.@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Cloud platform pricing Not available on partner-operated cloud platforms.@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Text editor tool@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN The text editor tool uses the same pricing structure as other tools used with Claude. It follows the standard input and output token pricing based on the Claude model you're using.@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN In addition to the base tokens, the following additional input tokens are needed for the text editor tool:@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Tool Additional input tokens@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN text_editor_20250429 (Claude 4.x) 700 tokens@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN See tool use pricing for complete pricing details.@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Session runtime@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN SKU Rate Metering@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Session runtime $0.08 per session-hour running status duration@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Runtime is measured to the millisecond and accrues only while the session's status is running. Time spent idle (waiting for your next message or a tool confirmation), rescheduling, or terminated does not count toward runtime.@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Web search tool@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Web search usage is charged in addition to token usage:@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN { "usage" : { "input_tokens" : 105 , "output_tokens" : 6039 , "cache_read_input_tokens" : 7123 , "cache_creation_input_tokens" : 7345 , "server_tool_use" : { "web_search_requests" : 1 } } }@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN \ue056@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Web search is available on the Claude API for $10 per 1,000 searches, plus standard token costs for search-generated content. Web search results retrieved throughout a conversation are counted as input tokens, in search iterations executed during a single turn and in subsequent conversation turns.@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Each web search counts as one use, regardless of the number of results returned. If an error occurs during web search, the web search will not be billed.@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN \ue09a@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Web fetch tool@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Web fetch usage has no additional charges beyond standard token costs:@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN { "usage" : { "input_tokens" : 25039 , "output_tokens" : 931 , "cache_read_input_tokens" : 0 , "cache_creation_input_tokens" : 0 , "server_tool_use" : { "web_fetch_requests" : 1 } } }@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN \ue056@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN The web fetch tool is available on the Claude API at no additional cost. You only pay standard token costs for the fetched content that becomes part of your conversation context.@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN To protect against inadvertently fetching large content that would consume excessive tokens, use the max_content_tokens parameter to set appropriate limits based on your use case and budget considerations.@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Example token usage for typical content:@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Average web page (10 kB): ~2,500 tokens@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Large documentation page (100 kB): ~25,000 tokens@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Research paper PDF (500 kB): ~125,000 tokens@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN \ue09a@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Computer use tool@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Computer use follows the standard tool use pricing. When using the computer use tool:@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN System prompt overhead: The computer use beta adds 466-499 tokens to the system prompt@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Computer use tool token usage:@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Model Input tokens per tool definition@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Claude 4.x models 735 tokens@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Additional token consumption:@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Screenshot images (see Vision pricing)@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Tool execution results returned to Claude@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Session runtime replaces the Code Execution container-hour billing model when using Claude Managed Agents. You are not separately billed for container hours on top of session runtime.@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN If you're also using bash or text editor tools alongside computer use, those tools have their own token costs as documented in their respective pages.@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Worked example@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN A one-hour coding session using Claude Opus 4.8 that consumes 50,000 input tokens and 15,000 output tokens:@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Line item Calculation Cost@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Input tokens 50,000 \u00d7 $5 / 1,000,000 $0.25@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Output tokens 15,000 \u00d7 $25 / 1,000,000 $0.375@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Session runtime 1.0 hour \u00d7 $0.08 $0.08@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Total $0.705@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN If prompt caching is active and 40,000 of the input tokens are cache reads:@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Line item Calculation Cost@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Uncached input tokens 10,000 \u00d7 $5 / 1,000,000 $0.05@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Cache read tokens 40,000 \u00d7 $5 \u00d7 0.1 / 1,000,000 $0.02@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Output tokens 15,000 \u00d7 $25 / 1,000,000 $0.375@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Session runtime 1.0 hour \u00d7 $0.08 $0.08@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Total $0.525@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Claude Managed Agents pricing@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Claude Managed Agents is billed on two dimensions: tokens and session runtime.@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN \ue09a@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Tokens@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN All tokens consumed by a Claude Managed Agents session are billed at the rates shown in Model pricing. Prompt caching multipliers apply identically. Web search triggered inside a session incurs the standard $10 per 1,000 searches. On Claude Platform on AWS, session token and runtime charges convert to Claude Consumption Units at the standard rate.@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN The following Messages API modifiers do not apply to Claude Managed Agents sessions:@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Modifier Why it doesn't apply@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Batch API discount Sessions are stateful and interactive. There is no batch mode.@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Fast mode premium Inference speed is managed by the runtime.@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Data residency multiplier inference_geo is a Messages API request field.@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Cloud platform pricing Not available on partner-operated cloud platforms.@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN \ue09a@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Session runtime@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN SKU Rate Metering@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Session runtime $0.08 per session-hour running status duration@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Runtime is measured to the millisecond and accrues only while the session's status is running. Time spent idle (waiting for your next message or a tool confirmation), rescheduling, or terminated does not count toward runtime.@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Example calculation for processing 10,000 support tickets:@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Average ~3,700 tokens per conversation@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Using Claude Haiku 4.5 at $1/MTok input, $5/MTok output@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Total cost: ~$37.00 per 10,000 tickets@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN For a detailed walkthrough of this calculation, see the customer support agent guide.@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Session runtime replaces the Code Execution container-hour billing model when using Claude Managed Agents. You are not separately billed for container hours on top of session runtime.@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Additional pricing considerations@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN \ue09a@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Cost optimization strategies@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN When building agents with Claude:@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN 1. Use appropriate models: Choose Haiku for simple tasks, Sonnet for most production workloads, and Opus for the most complex reasoning@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN 2. Implement prompt caching: Reduce costs for repeated context@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN 3. Batch operations: Use the Batch API for non-time-sensitive tasks@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN 4. Monitor usage patterns: Track token consumption to identify optimization opportunities@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN \ue097@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN For high-volume agent applications, contact the enterprise sales team for custom pricing arrangements.@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN \ue09a@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Rate limits@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Rate limits vary by usage tier and affect how many requests you can make:@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Start tier: Entry-level limits for getting started@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Build tier: Increased limits for growing applications@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Scale tier: Highest standard limits for production workloads@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN For detailed rate limit information, see Rate limits.@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN For limits beyond the Scale tier or custom pricing arrangements, contact the sales team.@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN \ue09a@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Volume discounts@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Volume discounts may be available for high-volume users. These are negotiated on a case-by-case basis.@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Standard usage tiers use the pricing shown in Model pricing@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Enterprise customers can contact sales for custom pricing@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Academic and research discounts may be available@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN \ue09a@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Enterprise pricing@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN For enterprise customers with specific needs:@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Custom rate limits@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Volume discounts@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Dedicated support@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Custom terms@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Contact the sales team at sales@anthropic.com or through the Claude Console to discuss enterprise pricing options.@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN \ue09a@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Billing and payment@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Billing is based on actual monthly usage@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * All payments are in USD@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Credit card and invoicing options available@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Usage tracking available in the Claude Console@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN \ue09a@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Frequently asked questions@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN How is token usage calculated?@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Tokens are pieces of text that models process. As a rough estimate, 1 token is approximately 4 characters or 0.75 words in English. The exact count varies by language and content type.@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Are there free tiers or trials?@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN New users receive a small amount of free credits to test the API. Contact sales for information about extended trials for enterprise evaluation.@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN How do discounts stack?@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Batch API and prompt caching discounts can be combined. For example, using both features together provides significant cost savings compared to standard API calls. See prompt caching pricing for how the multipliers interact.@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN What payment methods are accepted?@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Major credit cards are accepted for standard accounts. Enterprise customers can arrange invoicing and other payment methods.@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN For additional questions about pricing, contact support@anthropic.com.@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Was this page helpful?@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN \ue0fb \ue0f9@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Model pricing@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Cloud platform pricing@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Claude Platform on AWS pricing@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Inference geography@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Private offers@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Claude in Microsoft Foundry pricing@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Inference geography@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Feature-specific pricing@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Prompt caching@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Data residency pricing@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Fast mode pricing@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Batch processing@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Long context pricing@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Tool use pricing@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Specific tool pricing@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Claude Managed Agents pricing@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Tokens@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Session runtime@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Worked example@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Additional pricing considerations@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Cost optimization strategies@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Rate limits@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Volume discounts@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Enterprise pricing@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Billing and payment@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Frequently asked questions@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Claude Platform Docs@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Solutions@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * AI agents@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Code modernization@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Coding@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Customer support@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Education@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Financial services@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Government@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Life sciences@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Partners@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Claude on AWS@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Claude on Google Cloud@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Learn@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Blog@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Courses@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Use cases@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Connectors@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Customer stories@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Engineering at Anthropic@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Events@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Powered by Claude@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Service partners@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Startups program@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Company@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Anthropic@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Careers@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Economic Futures@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Research@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * News@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Responsible Scaling Policy@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Security and compliance@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Transparency@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Learn@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Blog@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Courses@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Use cases@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Connectors@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Customer stories@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Engineering at Anthropic@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Events@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Powered by Claude@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Service partners@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Startups program@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Help and security@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Availability@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Status@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Support@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Discord@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Terms and policies@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Privacy policy@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Responsible disclosure policy@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Terms of service: Commercial@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Terms of service: Consumer@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Usage policy@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Worked example@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN A one-hour coding session using Claude Opus 4.8 that consumes 50,000 input tokens and 15,000 output tokens:@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Line item Calculation Cost@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Input tokens 50,000 \u00d7 $5 / 1,000,000 $0.25@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Output tokens 15,000 \u00d7 $25 / 1,000,000 $0.375@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Session runtime 1.0 hour \u00d7 $0.08 $0.08@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Total $0.705@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN If prompt caching is active and 40,000 of the input tokens are cache reads:@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Line item Calculation Cost@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Uncached input tokens 10,000 \u00d7 $5 / 1,000,000 $0.05@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Cache read tokens 40,000 \u00d7 $5 \u00d7 0.1 / 1,000,000 $0.02@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Output tokens 15,000 \u00d7 $25 / 1,000,000 $0.375@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Session runtime 1.0 hour \u00d7 $0.08 $0.08@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Total $0.525@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN \ue08f@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Example calculation for processing 10,000 support tickets:@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Average ~3,700 tokens per conversation@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Using Claude Haiku 4.5 at $1/MTok input, $5/MTok output@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Total cost: ~$37.00 per 10,000 tickets@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN For a detailed walkthrough of this calculation, see the customer support agent guide.@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN \ue09a@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Additional pricing considerations@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN \ue09a@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Cost optimization strategies@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN When building agents with Claude:@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN 1. Use appropriate models: Choose Haiku for simple tasks, Sonnet for most production workloads, and Opus for the most complex reasoning@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN 2. Implement prompt caching: Reduce costs for repeated context@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN 3. Batch operations: Use the Batch API for non-time-sensitive tasks@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN 4. Monitor usage patterns: Track token consumption to identify optimization opportunities@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN \ue097@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN For high-volume agent applications, contact the enterprise sales team for custom pricing arrangements.@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN \ue09a@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Rate limits@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Rate limits vary by usage tier and affect how many requests you can make:@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Start tier: Entry-level limits for getting started@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Build tier: Increased limits for growing applications@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Scale tier: Highest standard limits for production workloads@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN For detailed rate limit information, see Rate limits.@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN For limits beyond the Scale tier or custom pricing arrangements, contact the sales team.@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN \ue09a@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Volume discounts@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Volume discounts may be available for high-volume users. These are negotiated on a case-by-case basis.@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Standard usage tiers use the pricing shown in Model pricing@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Enterprise customers can contact sales for custom pricing@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Academic and research discounts may be available@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN \ue09a@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Enterprise pricing@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN For enterprise customers with specific needs:@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Custom rate limits@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Volume discounts@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Dedicated support@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Custom terms@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Contact the sales team at sales@anthropic.com or through the Claude Console to discuss enterprise pricing options.@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN \ue09a@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Billing and payment@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Billing is based on actual monthly usage@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * All payments are in USD@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Credit card and invoicing options available@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Usage tracking available in the Claude Console@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN \ue09a@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Frequently asked questions@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN How is token usage calculated?@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Tokens are pieces of text that models process. As a rough estimate, 1 token is approximately 4 characters or 0.75 words in English. The exact count varies by language and content type.@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Are there free tiers or trials?@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN New users receive a small amount of free credits to test the API. Contact sales for information about extended trials for enterprise evaluation.@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN How do discounts stack?@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Batch API and prompt caching discounts can be combined. For example, using both features together provides significant cost savings compared to standard API calls. See prompt caching pricing for how the multipliers interact.@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN What payment methods are accepted?@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Major credit cards are accepted for standard accounts. Enterprise customers can arrange invoicing and other payment methods.@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN For additional questions about pricing, contact support@anthropic.com.@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Was this page helpful?@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN \ue0fb \ue0f9@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Model pricing@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Cloud platform pricing@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Claude Platform on AWS pricing@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Inference geography@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Private offers@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Claude in Microsoft Foundry pricing@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Inference geography@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Feature-specific pricing@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Prompt caching@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Data residency pricing@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Fast mode pricing@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Batch processing@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Long context pricing@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Tool use pricing@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Specific tool pricing@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Claude Managed Agents pricing@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Tokens@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Session runtime@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Worked example@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Additional pricing considerations@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Cost optimization strategies@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Rate limits@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Volume discounts@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Enterprise pricing@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Billing and payment@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Frequently asked questions@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Claude Platform Docs@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Solutions@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * AI agents@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Code modernization@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Coding@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Customer support@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Education@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Financial services@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Government@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Life sciences@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Partners@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Claude on AWS@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Claude on Google Cloud@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Learn@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Blog@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Courses@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Use cases@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Connectors@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Customer stories@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Engineering at Anthropic@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Events@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Powered by Claude@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Service partners@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Startups program@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Company@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Anthropic@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Careers@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Economic Futures@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Research@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * News@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Responsible Scaling Policy@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Security and compliance@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Transparency@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Learn@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Blog@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Courses@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Use cases@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Connectors@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Customer stories@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Engineering at Anthropic@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Events@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Powered by Claude@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Service partners@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Startups program@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Help and security@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Availability@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Status@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Support@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Discord@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Terms and policies@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Privacy policy@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Responsible disclosure policy@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Terms of service: Commercial@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Terms of service: Consumer@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Usage policy@changed_PLACEMARKER_CLOSED\n@changed_into_PLACEMARKER_OPEN * If files are included in the request, execution time is billed even if the tool is not invoked, due tobecause files beingare preloaded onto the container@changed_into_PLACEMARKER_CLOSED", "diff_clean": "", "diff_added": "", "diff_added_clean": "", "diff_full": "", "diff_full_clean": "", "diff_patch": "", "diff_removed": "", "diff_removed_clean": "", "diff_changed_from": "", "diff_changed_to": "", "diff_url": "https://changedetection.benfou.fr/diff/f10610bf-32e8-4762-a1c2-e9481624471d", "raw_diff": "@changed_PLACEMARKER_OPEN Cookie settings@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN We use cookies to deliver and improve our services, analyze site usage, and if you agree, to customize or personalize your experience and market our services to you. You can read our Cookie Policy here.@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Customize Customize Cookie Settings Reject Reject All Cookies Accept Accept All Cookies@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Claude Platform Docs@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Messages@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Managed Agents@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Admin@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Resources \ue027@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN \ue04a API reference@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN \ue053 English \ue027 \ue04f ConsoleLog in@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN \ue0d3 \ue062@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN \ue0d3@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Search...@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN \u2318K@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Models@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Models overview Model IDs and versioning Choosing a model Introducing Claude Fable 5 and Claude Mythos 5 What's new in Claude Opus 4.8 What's new in Claude Sonnet 5 Upgrade between model versions Model deprecations Model cards System prompts Pricing@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN \ue104@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Log in@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN \ue0dd Models & pricing \ue027 Pricing@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Models & pricing / Models@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Pricing@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Cookie settings@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN We use cookies to deliver and improve our services, analyze site usage, and if you agree, to customize or personalize your experience and market our services to you. You can read our Cookie Policy here.@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Customize Customize Cookie Settings Reject Reject All Cookies Accept Accept All Cookies@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Claude Platform Docs@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Messages@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Managed Agents@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Admin@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Resources \ue027@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN \ue04a API reference@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN \ue053 English \ue027 \ue04f ConsoleLog in@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN \ue0d3 \ue062@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN \ue0d3@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Search...@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN \u2318K@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Models@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Models overview Model IDs and versioning Choosing a model Introducing Claude Fable 5 and Claude Mythos 5 What's new in Claude Opus 4.8 What's new in Claude Sonnet 5 Upgrade between model versions Model deprecations Model cards System prompts Pricing@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN \ue104@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Log in@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN \ue0dd Models & pricing \ue027 Pricing@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Models & pricing / Models@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Pricing@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Copy page@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN \ue027@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Learn about Anthropic's pricing structure for models and features@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Learn about Anthropic's pricing structure for models and features@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Copy page@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN \ue027@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN This page provides detailed pricing information for Anthropic's models and features. All prices are in USD.@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN For the most current pricing information, visit claude.com/pricing.@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN \ue09a@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Model pricing@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN The following table shows pricing for all Claude models:@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Model Base Input Tokens 5m Cache Writes 1h Cache Writes Cache Hits & Refreshes Output Tokens@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Claude Fable 5 $10 / MTok $12.50 / MTok $20 / MTok $1 / MTok $50 / MTok@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Claude Mythos 5 (limited availability) $10 / MTok $12.50 / MTok $20 / MTok $1 / MTok $50 / MTok@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Claude Opus 4.8 $5 / MTok $6.25 / MTok $10 / MTok $0.50 / MTok $25 / MTok@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Claude Opus 4.7 $5 / MTok $6.25 / MTok $10 / MTok $0.50 / MTok $25 / MTok@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Claude Opus 4.6 $5 / MTok $6.25 / MTok $10 / MTok $0.50 / MTok $25 / MTok@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Claude Opus 4.5 $5 / MTok $6.25 / MTok $10 / MTok $0.50 / MTok $25 / MTok@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Claude Opus 4.1 (deprecated) $15 / MTok $18.75 / MTok $30 / MTok $1.50 / MTok $75 / MTok@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Claude Opus 4 (retired, except on Google Cloud) $15 / MTok $18.75 / MTok $30 / MTok $1.50 / MTok $75 / MTok@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Claude Sonnet 5 $2 / MTok $2.50 / MTok $4 / MTok $0.20 / MTok $10 / MTok@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN through August 31, 2026@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Claude Sonnet 5 $3 / MTok $3.75 / MTok $6 / MTok $0.30 / MTok $15 / MTok@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN starting September 1, 2026@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Claude Sonnet 4.6 $3 / MTok $3.75 / MTok $6 / MTok $0.30 / MTok $15 / MTok@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Claude Sonnet 4.5 $3 / MTok $3.75 / MTok $6 / MTok $0.30 / MTok $15 / MTok@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Claude Sonnet 4 (retired, except on Bedrock and Google Cloud) $3 / MTok $3.75 / MTok $6 / MTok $0.30 / MTok $15 / MTok@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Claude Haiku 4.5 $1 / MTok $1.25 / MTok $2 / MTok $0.10 / MTok $5 / MTok@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Claude Haiku 3.5 (retired, except on Bedrock and Google Cloud) $0.80 / MTok $1 / MTok $1.60 / MTok $0.08 / MTok $4 / MTok@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN \ue08f@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Introductory pricing of $2/$10 per million input/output tokens is in effect through August 31, 2026, after which the standard pricing of $3/$15 per million input/output tokens will take effect.@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN \ue08f@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN MTok = Million tokens. The "Base Input Tokens" column shows standard input pricing, the "5m Cache Writes", "1h Cache Writes", and "Cache Hits & Refreshes" columns are specific to prompt caching, and "Output Tokens" shows output pricing. See prompt caching pricing for an explanation of the cache columns and pricing multipliers.@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN \ue08f@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Claude Opus 4.7 and later Opus models, Claude Fable 5, Claude Mythos 5, Claude Mythos Preview, and Claude Sonnet 5 use a newer tokenizer that contributes to their improved performance on a wide range of tasks. This tokenizer produces approximately 30% more tokens for the same text. The exact increase depends on the content and workload shape. Claude Sonnet 4.6 and earlier models use the previous tokenizer.@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN For Claude Platform on AWS pricing, see Claude Platform on AWS pricing.@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN This page provides detailed pricing information for Anthropic's models and features. All prices are in USD.@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN For the most current pricing information, visit claude.com/pricing.@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Cloud platform pricing@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN This section covers partner-operated cloud platforms, where the cloud provider invoices you. For Anthropic-operated cloud platforms billed through a marketplace, see Claude Platform on AWS pricing and Claude in Microsoft Foundry pricing.@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Claude models are available on Amazon Bedrock and Google Cloud. For official pricing, visit:@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Amazon Bedrock pricing@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Google Cloud pricing@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN \ue08f@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Regional and multi-region endpoint pricing for Claude 4.5 models and beyond@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Starting with Claude Sonnet 4.5, Haiku 4.5, and Opus 4.5:@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Bedrock offers two endpoint types: global endpoints (dynamic routing for maximum availability) and regional endpoints (guaranteed data routing through specific geographic regions).@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Google Cloud offers three endpoint types: global endpoints, multi-region endpoints (dynamic routing within a geographic area), and regional endpoints.@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Regional and multi-region endpoints include a 10% premium over global endpoints. The Claude API (first-party) is global by default; for first-party data residency options and pricing, see Data residency pricing.@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Scope: This pricing structure applies to Claude Sonnet 4.5, Haiku 4.5, Opus 4.5, and all future models. Earlier models (Claude Opus 4.1 (deprecated) and prior releases) retain their existing pricing.@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN For implementation details and code examples:@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Amazon Bedrock global vs regional endpoints for Opus 4.7, Haiku 4.5, and later models, or the legacy integration for all other models on Bedrock@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Google Cloud global, multi-region, and regional endpoints@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN \ue09a@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Claude Platform on AWS pricing@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Claude Platform on AWS bills through AWS Marketplace using Claude Consumption Units (CCUs). Anthropic rates your token usage in USD at standard per-model, per-feature rates, applies any negotiated discount, converts the result to CCUs at $0.01 per CCU, and reports the CCU quantity to AWS Marketplace hourly. Your AWS bill shows a single CCU line item.@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Concept Details@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Billing unit Claude Consumption Unit (CCU)@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN CCU price $0.01 per CCU (fixed; discounts apply at token-to-CCU conversion, not to the CCU price)@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Conversion Token usage rated in USD at standard per-model, per-feature rates (same as Claude API pricing), then converted to CCUs at $0.01 per CCU@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Billing cadence Hourly metering to AWS Marketplace; monthly invoices@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Payment model Arrears only (postpaid); no prepaid credits@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Discounts Applied as fewer CCUs metered@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Tax Pre-tax metering; AWS Marketplace handles tax@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Cost visibility Real-time breakdown in the Claude Console (access through the AWS Console); AWS Cost Explorer shows aggregated CCU@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Model pricing@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN The following table shows pricing for all Claude models:@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Model Base Input Tokens 5m Cache Writes 1h Cache Writes Cache Hits & Refreshes Output Tokens@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Claude Fable 5 $10 / MTok $12.50 / MTok $20 / MTok $1 / MTok $50 / MTok@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Claude Mythos 5 (limited availability) $10 / MTok $12.50 / MTok $20 / MTok $1 / MTok $50 / MTok@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Claude Opus 4.8 $5 / MTok $6.25 / MTok $10 / MTok $0.50 / MTok $25 / MTok@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Claude Opus 4.7 $5 / MTok $6.25 / MTok $10 / MTok $0.50 / MTok $25 / MTok@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Claude Opus 4.6 $5 / MTok $6.25 / MTok $10 / MTok $0.50 / MTok $25 / MTok@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Claude Opus 4.5 $5 / MTok $6.25 / MTok $10 / MTok $0.50 / MTok $25 / MTok@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Claude Opus 4.1 (deprecated) $15 / MTok $18.75 / MTok $30 / MTok $1.50 / MTok $75 / MTok@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Claude Opus 4 (retired, except on Google Cloud) $15 / MTok $18.75 / MTok $30 / MTok $1.50 / MTok $75 / MTok@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Claude Sonnet 5 $2 / MTok $2.50 / MTok $4 / MTok $0.20 / MTok $10 / MTok@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN through August 31, 2026@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Claude Sonnet 5 $3 / MTok $3.75 / MTok $6 / MTok $0.30 / MTok $15 / MTok@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN starting September 1, 2026@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Claude Sonnet 4.6 $3 / MTok $3.75 / MTok $6 / MTok $0.30 / MTok $15 / MTok@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Claude Sonnet 4.5 $3 / MTok $3.75 / MTok $6 / MTok $0.30 / MTok $15 / MTok@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Claude Sonnet 4 (retired, except on Bedrock and Google Cloud) $3 / MTok $3.75 / MTok $6 / MTok $0.30 / MTok $15 / MTok@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Claude Haiku 4.5 $1 / MTok $1.25 / MTok $2 / MTok $0.10 / MTok $5 / MTok@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Claude Haiku 3.5 (retired, except on Bedrock and Google Cloud) $0.80 / MTok $1 / MTok $1.60 / MTok $0.08 / MTok $4 / MTok@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Claude Consumption Units. If Customer accesses the Services through certain Marketplace Platforms (e.g., Claude Platform on AWS), usage will be invoiced in Claude Consumption Units ("CCU") rather than per MTok. A CCU is a unit of measure used solely for Marketplace Platform invoicing. One hundred (100) CCU represents $1.00 USD of fees owed for the Services, calculated at the applicable prices on claude.com/pricing#api, after application of any discounts.@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Introductory pricing of $2/$10 per million input/output tokens is in effect through August 31, 2026, after which the standard pricing of $3/$15 per million input/output tokens will take effect.@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN \ue08f@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN MTok = Million tokens. The "Base Input Tokens" column shows standard input pricing, the "5m Cache Writes", "1h Cache Writes", and "Cache Hits & Refreshes" columns are specific to prompt caching, and "Output Tokens" shows output pricing. See prompt caching pricing for an explanation of the cache columns and pricing multipliers.@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN \ue08f@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Claude Opus 4.7 and later Opus models, Claude Fable 5, Claude Mythos 5, Claude Mythos Preview, and Claude Sonnet 5 use a newer tokenizer that contributes to their improved performance on a wide range of tasks. This tokenizer produces approximately 30% more tokens for the same text. The exact increase depends on the content and workload shape. Claude Sonnet 4.6 and earlier models use the previous tokenizer.@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN For Claude Platform on AWS pricing, see Claude Platform on AWS pricing.@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Inference geography@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN For Claude Opus 4.6, Claude Sonnet 4.6, and later models, using inference_geo: "us" applies a 1.1x pricing multiplier. inference_geo: "global" (default) uses standard pricing. See Data residency for details.@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Cloud platform pricing@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN This section covers partner-operated cloud platforms, where the cloud provider invoices you. For Anthropic-operated cloud platforms billed through a marketplace, see Claude Platform on AWS pricing and Claude in Microsoft Foundry pricing.@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Claude models are available on Amazon Bedrock and Google Cloud. For official pricing, visit:@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Amazon Bedrock pricing@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Google Cloud pricing@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN \ue08f@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Regional and multi-region endpoint pricing for Claude 4.5 models and beyond@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Starting with Claude Sonnet 4.5, Haiku 4.5, and Opus 4.5:@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Bedrock offers two endpoint types: global endpoints (dynamic routing for maximum availability) and regional endpoints (guaranteed data routing through specific geographic regions).@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Google Cloud offers three endpoint types: global endpoints, multi-region endpoints (dynamic routing within a geographic area), and regional endpoints.@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Regional and multi-region endpoints include a 10% premium over global endpoints. The Claude API (first-party) is global by default; for first-party data residency options and pricing, see Data residency pricing.@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Scope: This pricing structure applies to Claude Sonnet 4.5, Haiku 4.5, Opus 4.5, and all future models. Earlier models (Claude Opus 4.1 (deprecated) and prior releases) retain their existing pricing.@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN For implementation details and code examples:@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Amazon Bedrock global vs regional endpoints for Opus 4.7, Haiku 4.5, and later models, or the legacy integration for all other models on Bedrock@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Google Cloud global, multi-region, and regional endpoints@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Private offers@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN When you sign up on the AWS Console Claude Platform on AWS service page, the AWS Console looks up any private offer associated with your account and prompts you to accept it in AWS Marketplace. Contact your Anthropic account representative for private offer terms.@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN \ue08f@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN If you have an existing Amazon Bedrock private offer, contact your Anthropic or AWS account representative before getting started with Claude Platform on AWS to ensure your discounts are applied correctly. Discounts cannot be applied retroactively to usage incurred before your private offer is accepted.@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN \ue09a@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Claude in Microsoft Foundry pricing@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Claude in Microsoft Foundry bills through the Azure Marketplace using Claude Consumption Units (CCUs). Anthropic rates your token usage in USD at standard per-model, per-feature rates, applies any negotiated discount, converts the result to CCUs at $0.01 per CCU, and reports the CCU quantity to the Azure Marketplace hourly. Your Azure bill shows a single CCU line item.@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Claude Platform on AWS pricing@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Claude Platform on AWS bills through AWS Marketplace using Claude Consumption Units (CCUs). Anthropic rates your token usage in USD at standard per-model, per-feature rates, applies any negotiated discount, converts the result to CCUs at $0.01 per CCU, and reports the CCU quantity to AWS Marketplace hourly. Your AWS bill shows a single CCU line item.@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Billing cadence Hourly metering to the AzureAWS Marketplace; monthly invoices@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Tax Pre-tax metering; Azure Marketplace handles tax@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Cost visibility Azure Cost Management shows aggregated CCU@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Tax Pre-tax metering; AWS Marketplace handles tax@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Cost visibility Real-time breakdown in the Claude Console (access through the AWS Console); AWS Cost Explorer shows aggregated CCU@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Claude Consumption Units. If Customer accesses the Services through certain Marketplace Platforms (e.g., Claude Platform on AWS, Claude in Microsoft Foundry),AWS), usage will be invoiced in Claude Consumption Units ("CCU") rather than per MTok. A CCU is a unit of measure used solely for Marketplace Platform invoicing. One hundred (100) CCU represents $1.00 USD of fees owed for the Services, calculated at the applicable prices on claude.com/pricing#api, after application of any discounts.@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Deployments hosted on Azure canFor useClaude theOpus US Data4.6, Zone Standard deployment type, which keeps inference within theClaude UnitedSonnet States.4.6, Thisand islater equivalentmodels, tousing inference_geo: "us" onapplies thea Claude1.1x APIpricing andmultiplier. appliesinference_geo: the"global" same(default) 1.1xuses pricingstandard multiplier.pricing. See Data residency for details.@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Feature-specific pricing@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Private offers@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN When you sign up on the AWS Console Claude Platform on AWS service page, the AWS Console looks up any private offer associated with your account and prompts you to accept it in AWS Marketplace. Contact your Anthropic account representative for private offer terms.@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN \ue08f@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN If you have an existing Amazon Bedrock private offer, contact your Anthropic or AWS account representative before getting started with Claude Platform on AWS to ensure your discounts are applied correctly. Discounts cannot be applied retroactively to usage incurred before your private offer is accepted.@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Prompt caching@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Prompt caching reduces costs and latency by reusing previously processed portions of your prompt across API calls. Instead of reprocessing the same large system prompt, document, or conversation history on every request, the API reads from cache at a fraction of the standard input price.@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN There are two ways to enable prompt caching:@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Automatic caching: Add a single cache_control field at the top level of your request. The system automatically manages cache breakpoints as conversations grow. This is the recommended starting point for most use cases.@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Explicit cache breakpoints: Place cache_control directly on individual content blocks for fine-grained control over exactly what gets cached.@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Prompt caching uses the following pricing multipliers relative to base input token rates:@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Cache operation Multiplier Duration@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN 5-minute cache write 1.25x base input price Cache valid for 5 minutes@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN 1-hour cache write 2x base input price Cache valid for 1 hour@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Cache read (hit) 0.1x base input price Same duration as the preceding write@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Cache write tokens are charged when content is first stored. Cache read tokens are charged when a subsequent request retrieves the cached content. A cache hit costs 10% of the standard input price, which means caching pays off after just one cache read for the 5-minute duration (1.25x write), or after two cache reads for the 1-hour duration (2x write).@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN These multipliers stack with other pricing modifiers, including the Batch API discount and data residency.@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN For implementation details, supported models, and code examples, see Prompt caching.@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Claude in Microsoft Foundry pricing@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Claude in Microsoft Foundry bills through the Azure Marketplace using Claude Consumption Units (CCUs). Anthropic rates your token usage in USD at standard per-model, per-feature rates, applies any negotiated discount, converts the result to CCUs at $0.01 per CCU, and reports the CCU quantity to the Azure Marketplace hourly. Your Azure bill shows a single CCU line item.@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Concept Details@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Billing unit Claude Consumption Unit (CCU)@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN CCU price $0.01 per CCU (fixed; discounts apply at token-to-CCU conversion, not to the CCU price)@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Conversion Token usage rated in USD at standard per-model, per-feature rates (same as Claude API pricing), then converted to CCUs at $0.01 per CCU@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Billing cadence Hourly metering to the Azure Marketplace; monthly invoices@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Payment model Arrears only (postpaid); no prepaid credits@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Discounts Applied as fewer CCUs metered@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Tax Pre-tax metering; Azure Marketplace handles tax@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Cost visibility Azure Cost Management shows aggregated CCU@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN \ue08f@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Claude Consumption Units. If Customer accesses the Services through certain Marketplace Platforms (e.g., Claude Platform on AWS, Claude in Microsoft Foundry), usage will be invoiced in Claude Consumption Units ("CCU") rather than per MTok. A CCU is a unit of measure used solely for Marketplace Platform invoicing. One hundred (100) CCU represents $1.00 USD of fees owed for the Services, calculated at the applicable prices on claude.com/pricing#api, after application of any discounts.@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Data residency pricing@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN For Claude Opus 4.6, Claude Sonnet 4.6, and later models, specifying US-only inference through the inference_geo parameter incurs a 1.1x multiplier on all token pricing categories, including input tokens, output tokens, cache writes, and cache reads. Global routing (the default) uses standard pricing.@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN This applies to the Claude API (first-party) and Claude Platform on AWS. On Claude in Microsoft Foundry, the same 1.1x multiplier applies to deployments that use the US Data Zone Standard deployment type (see Inference geography). Partner-operated platforms (Bedrock and Google Cloud) have independent regional pricing. See Bedrock and Google Cloud for details. Earlier models do not support the inference_geo parameter and always use standard pricing; requests that include the parameter on these models return a 400 error.@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN For more information, see Data residency.@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Inference geography@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Deployments hosted on Azure can use the US Data Zone Standard deployment type, which keeps inference within the United States. This is equivalent to inference_geo: "us" on the Claude API and applies the same 1.1x pricing multiplier. See Data residency for details.@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Fast mode pricing@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Fast mode, in research preview, provides significantly faster output for Claude Opus 4.8 and Claude Opus 4.7 at premium pricing. Fast mode pricing applies across the full context window, including requests over 200k input tokens. Fast mode is not available on Claude Platform on AWS.@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Model Input Output@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Claude Opus 4.8 $10 / MTok $50 / MTok@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Claude Opus 4.7 $30 / MTok $150 / MTok@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Fast mode for Claude Opus 4.7 is deprecated and will be removed on July 24, 2026. As of June 29, 2026, fast mode is not available on Claude Opus 4.6: requests to claude-opus-4-6 with speed: "fast" run at standard speed and are billed at standard rates. See Fast mode.@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Fast mode pricing stacks with other pricing modifiers:@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Prompt caching multipliers apply on top of fast mode pricing@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Data residency multipliers apply on top of fast mode pricing@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Fast mode is not available with the Batch API.@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN For more information, see Fast mode.@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Feature-specific pricing@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN \ue09a@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Prompt caching@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Prompt caching reduces costs and latency by reusing previously processed portions of your prompt across API calls. Instead of reprocessing the same large system prompt, document, or conversation history on every request, the API reads from cache at a fraction of the standard input price.@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN There are two ways to enable prompt caching:@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Automatic caching: Add a single cache_control field at the top level of your request. The system automatically manages cache breakpoints as conversations grow. This is the recommended starting point for most use cases.@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Explicit cache breakpoints: Place cache_control directly on individual content blocks for fine-grained control over exactly what gets cached.@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Prompt caching uses the following pricing multipliers relative to base input token rates:@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Cache operation Multiplier Duration@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN 5-minute cache write 1.25x base input price Cache valid for 5 minutes@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN 1-hour cache write 2x base input price Cache valid for 1 hour@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Cache read (hit) 0.1x base input price Same duration as the preceding write@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Cache write tokens are charged when content is first stored. Cache read tokens are charged when a subsequent request retrieves the cached content. A cache hit costs 10% of the standard input price, which means caching pays off after just one cache read for the 5-minute duration (1.25x write), or after two cache reads for the 1-hour duration (2x write).@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN These multipliers stack with other pricing modifiers, including the Batch API discount and data residency.@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN For implementation details, supported models, and code examples, see Prompt caching.@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Batch processing@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN The Batch API allows asynchronous processing of large volumes of requests with a 50% discount on both input and output tokens.@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Model Batch input Batch output@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Claude Fable 5 $5 / MTok $25 / MTok@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Claude Mythos 5 (limited availability) $5 / MTok $25 / MTok@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Claude Opus 4.8 $2.50 / MTok $12.50 / MTok@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Claude Opus 4.7 $2.50 / MTok $12.50 / MTok@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Claude Opus 4.6 $2.50 / MTok $12.50 / MTok@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Claude Opus 4.5 $2.50 / MTok $12.50 / MTok@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Claude Opus 4.1 (deprecated) $7.50 / MTok $37.50 / MTok@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Claude Opus 4 (retired, except on Google Cloud) $7.50 / MTok $37.50 / MTok@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Claude Sonnet 5 $1 / MTok $5 / MTok@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN through August 31, 2026@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Claude Sonnet 5 $1.50 / MTok $7.50 / MTok@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN starting September 1, 2026@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Claude Sonnet 4.6 $1.50 / MTok $7.50 / MTok@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Claude Sonnet 4.5 $1.50 / MTok $7.50 / MTok@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Claude Sonnet 4 (retired, except on Bedrock and Google Cloud) $1.50 / MTok $7.50 / MTok@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Claude Haiku 4.5 $0.50 / MTok $2.50 / MTok@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Claude Haiku 3.5 (retired, except on Bedrock and Google Cloud) $0.40 / MTok $2 / MTok@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN For more information about batch processing, see Batch processing.@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Data residency pricing@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN For Claude Opus 4.6, Claude Sonnet 4.6, and later models, specifying US-only inference through the inference_geo parameter incurs a 1.1x multiplier on all token pricing categories, including input tokens, output tokens, cache writes, and cache reads. Global routing (the default) uses standard pricing.@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN This applies to the Claude API (first-party) and Claude Platform on AWS. On Claude in Microsoft Foundry, the same 1.1x multiplier applies to deployments that use the US Data Zone Standard deployment type (see Inference geography). Partner-operated platforms (Bedrock and Google Cloud) have independent regional pricing. See Bedrock and Google Cloud for details. Earlier models do not support the inference_geo parameter and always use standard pricing; requests that include the parameter on these models return a 400 error.@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN For more information, see Data residency.@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN \ue09a@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Fast mode pricing@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Fast mode, in research preview, provides significantly faster output for Claude Opus 4.8 and Claude Opus 4.7 at premium pricing. Fast mode pricing applies across the full context window, including requests over 200k input tokens. Fast mode is not available on Claude Platform on AWS.@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Model Input Output@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Claude Opus 4.8 $10 / MTok $50 / MTok@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Claude Opus 4.7 $30 / MTok $150 / MTok@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Fast mode for Claude Opus 4.7 is deprecated and will be removed on July 24, 2026. As of June 29, 2026, fast mode is not available on Claude Opus 4.6: requests to claude-opus-4-6 with speed: "fast" run at standard speed and are billed at standard rates. See Fast mode.@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Fast mode pricing stacks with other pricing modifiers:@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Prompt caching multipliers apply on top of fast mode pricing@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Data residency multipliers apply on top of fast mode pricing@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Fast mode is not available with the Batch API.@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN For more information, see Fast mode.@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Long context pricing@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Claude Fable 5, Claude Mythos 5, Claude Mythos Preview, Claude Opus 4.8, Opus 4.7, Opus 4.6, Sonnet 5, and Sonnet 4.6 include the full 1M token context window at standard pricing. (A 900k-token request is billed at the same per-token rate as a 9k-token request.) Prompt caching and batch processing discounts apply at standard rates across the full context window.@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN \ue09a@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Tool use pricing@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Tool use requests are priced based on:@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN 1. The total number of input tokens sent to the model (including in the tools parameter)@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN 2. The number of output tokens generated@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN 3. For server-side tools, additional usage-based pricing (e.g., web search charges per search performed)@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Client-side tools are priced the same as any other Claude API request, while server-side tools may incur additional charges based on their specific usage.@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN The additional tokens from tool use come from:@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * The tools parameter in API requests (tool names, descriptions, and schemas)@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * tool_use content blocks in API requests and responses@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * tool_result content blocks in API requests@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN When you use tools, the API also automatically includes a special system prompt for the model which enables tool use. The number of tool use tokens required for each model are listed below (excluding the additional tokens listed above). Note that the table assumes at least 1 tool is provided. If no tools are provided, then a tool choice of none uses 0 additional system prompt tokens.@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Model Tool choice Tool use system prompt token count@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Claude Opus 4.8 auto, noneany, tool 290 tokens410 tokens@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Claude Opus 4.7 auto, noneany, tool 675 tokens804 tokens@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Claude Opus 4.6 auto, noneany, tool 497 tokens589 tokens@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Claude Opus 4.5 auto, noneany, tool 496 tokens588 tokens@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Claude Opus 4.1 (deprecated) auto, noneany, tool 313 tokens315 tokens@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Claude Opus 4 (retired, except on Google Cloud) auto, noneany, tool 313 tokens315 tokens@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Claude Sonnet 5 auto, noneany, tool 354 tokens474 tokens@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Claude Sonnet 4.6 auto, noneany, tool 497 tokens589 tokens@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Claude Sonnet 4.5 auto, noneany, tool 496 tokens588 tokens@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Claude Sonnet 4 (retired, except on Bedrock and Google Cloud) auto, noneany, tool 313 tokens315 tokens@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Claude Haiku 4.5 auto, noneany, tool 496 tokens588 tokens@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Claude Haiku 3.5 (retired, except on Bedrock and Google Cloud) auto, noneany, tool 264 tokens355 tokens@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN These token counts are added to your normal input and output tokens to calculate the total cost of a request.@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN For current per-model prices, refer to the model pricing section.@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN For more information about tool use implementation and best practices, see Tool use.@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Batch processing@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN The Batch API allows asynchronous processing of large volumes of requests with a 50% discount on both input and output tokens.@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Model Batch input Batch output@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Claude Fable 5 $5 / MTok $25 / MTok@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Claude Mythos 5 (limited availability) $5 / MTok $25 / MTok@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Claude Opus 4.8 $2.50 / MTok $12.50 / MTok@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Claude Opus 4.7 $2.50 / MTok $12.50 / MTok@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Claude Opus 4.6 $2.50 / MTok $12.50 / MTok@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Claude Opus 4.5 $2.50 / MTok $12.50 / MTok@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Claude Opus 4.1 (deprecated) $7.50 / MTok $37.50 / MTok@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Claude Opus 4 (retired, except on Google Cloud) $7.50 / MTok $37.50 / MTok@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Claude Sonnet 5 $1 / MTok $5 / MTok@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN through August 31, 2026@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Claude Sonnet 5 $1.50 / MTok $7.50 / MTok@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN starting September 1, 2026@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Claude Sonnet 4.6 $1.50 / MTok $7.50 / MTok@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Claude Sonnet 4.5 $1.50 / MTok $7.50 / MTok@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Claude Sonnet 4 (retired, except on Bedrock and Google Cloud) $1.50 / MTok $7.50 / MTok@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Claude Haiku 4.5 $0.50 / MTok $2.50 / MTok@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Claude Haiku 3.5 (retired, except on Bedrock and Google Cloud) $0.40 / MTok $2 / MTok@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN For more information about batch processing, see Batch processing.@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Specific tool pricing@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Long context pricing@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Claude Fable 5, Claude Mythos 5, Claude Mythos Preview, Claude Opus 4.8, Opus 4.7, Opus 4.6, Sonnet 5, and Sonnet 4.6 include the full 1M token context window at standard pricing. (A 900k-token request is billed at the same per-token rate as a 9k-token request.) Prompt caching and batch processing discounts apply at standard rates across the full context window.@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Bash tool@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN The bash tool adds 245 input tokens to your API calls.@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Additional tokens are consumed by:@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Command outputs (stdout/stderr)@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Error messages@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Large file contents@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN See tool use pricing for complete pricing details.@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN \ue09a@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Code execution tool@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Code execution is free when used with web search or web fetch. When web_search_20260209 (or later) or web_fetch_20260209 (or later) is included in your API request, there are no additional charges for code execution tool calls beyond the standard input and output token costs.@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN When used without these tools, code execution is billed by execution time, tracked separately from token usage:@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Execution time has a minimum of 5 minutes@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Each organization receives 1,550 free hours of usage per month@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Additional usage beyond 1,550 hours is billed at $0.05 per hour, per container@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * If files are included in the request, execution time is billed even if the tool is not invoked, due to files being preloaded onto the container@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Code execution usage is tracked in the response:@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN { "usage" : { "input_tokens" : 105 , "output_tokens" : 239 , "server_tool_use" : { "code_execution_requests" : 1 } } }@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN \ue056@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN \ue09a@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Text editor tool@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN The text editor tool uses the same pricing structure as other tools used with Claude. It follows the standard input and output token pricing based on the Claude model you're using.@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN In addition to the base tokens, the following additional input tokens are needed for the text editor tool:@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Tool Additional input tokens@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN text_editor_20250429 (Claude 4.x) 700 tokens@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN See tool use pricing for complete pricing details.@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Tool use pricing@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Tool use requests are priced based on:@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN 1. The total number of input tokens sent to the model (including in the tools parameter)@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN 2. The number of output tokens generated@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN 3. For server-side tools, additional usage-based pricing (e.g., web search charges per search performed)@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Client-side tools are priced the same as any other Claude API request, while server-side tools may incur additional charges based on their specific usage.@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN The additional tokens from tool use come from:@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * The tools parameter in API requests (tool names, descriptions, and schemas)@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * tool_use content blocks in API requests and responses@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * tool_result content blocks in API requests@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN When you use tools, the API also automatically includes a special system prompt for the model which enables tool use. The number of tool use tokens required for each model are listed below (excluding the additional tokens listed above). Note that the table assumes at least 1 tool is provided. If no tools are provided, then a tool choice of none uses 0 additional system prompt tokens.@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Model Tool choice Tool use system prompt token count@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Claude Opus 4.8 auto, noneany, tool 290 tokens410 tokens@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Claude Opus 4.7 auto, noneany, tool 675 tokens804 tokens@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Claude Opus 4.6 auto, noneany, tool 497 tokens589 tokens@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Claude Opus 4.5 auto, noneany, tool 496 tokens588 tokens@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Claude Opus 4.1 (deprecated) auto, noneany, tool 313 tokens315 tokens@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Claude Opus 4 (retired, except on Google Cloud) auto, noneany, tool 313 tokens315 tokens@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Claude Sonnet 5 auto, noneany, tool 354 tokens474 tokens@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Claude Sonnet 4.6 auto, noneany, tool 497 tokens589 tokens@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Claude Sonnet 4.5 auto, noneany, tool 496 tokens588 tokens@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Claude Sonnet 4 (retired, except on Bedrock and Google Cloud) auto, noneany, tool 313 tokens315 tokens@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Claude Haiku 4.5 auto, noneany, tool 496 tokens588 tokens@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Claude Haiku 3.5 (retired, except on Bedrock and Google Cloud) auto, noneany, tool 264 tokens355 tokens@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN These token counts are added to your normal input and output tokens to calculate the total cost of a request.@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN For current per-model prices, refer to the model pricing section.@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN For more information about tool use implementation and best practices, see Tool use.@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Web search tool@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Web search usage is charged in addition to token usage:@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN { "usage" : { "input_tokens" : 105 , "output_tokens" : 6039 , "cache_read_input_tokens" : 7123 , "cache_creation_input_tokens" : 7345 , "server_tool_use" : { "web_search_requests" : 1 } } }@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN \ue056@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Web search is available on the Claude API for $10 per 1,000 searches, plus standard token costs for search-generated content. Web search results retrieved throughout a conversation are counted as input tokens, in search iterations executed during a single turn and in subsequent conversation turns.@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Each web search counts as one use, regardless of the number of results returned. If an error occurs during web search, the web search will not be billed.@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Specific tool pricing@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Web fetch tool@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Web fetch usage has no additional charges beyond standard token costs:@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN { "usage" : { "input_tokens" : 25039 , "output_tokens" : 931 , "cache_read_input_tokens" : 0 , "cache_creation_input_tokens" : 0 , "server_tool_use" : { "web_fetch_requests" : 1 } } }@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN \ue056@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN The web fetch tool is available on the Claude API at no additional cost. You only pay standard token costs for the fetched content that becomes part of your conversation context.@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN To protect against inadvertently fetching large content that would consume excessive tokens, use the max_content_tokens parameter to set appropriate limits based on your use case and budget considerations.@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Example token usage for typical content:@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Average web page (10 kB): ~2,500 tokens@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Large documentation page (100 kB): ~25,000 tokens@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Research paper PDF (500 kB): ~125,000 tokens@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN \ue09a@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Computer use tool@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Computer use follows the standard tool use pricing. When using the computer use tool:@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN System prompt overhead: The computer use beta adds 466-499 tokens to the system prompt@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Computer use tool token usage:@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Model Input tokens per tool definition@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Claude 4.x models 735 tokens@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Additional token consumption:@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Screenshot images (see Vision pricing)@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Tool execution results returned to Claude@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN \ue08f@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN If you're also using bash or text editor tools alongside computer use, those tools have their own token costs as documented in their respective pages.@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Bash tool@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN The bash tool definition adds the following input tokens to your request. This is in addition to the per-model tool use system prompt that applies whenever any tool is present.@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Model Additional input tokens@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Claude Opus 4.7 and Claude Opus 4.8 325 tokens@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Claude Opus 4.6, Claude Sonnet 4.6, and earlier 244 tokens@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Additional tokens are consumed by:@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Command outputs (stdout/stderr)@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Error messages@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Large file contents@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN See tool use pricing for complete pricing details.@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Claude Managed Agents pricing@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Claude Managed Agents is billed on two dimensions: tokens and session runtime.@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Code execution tool@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Code execution is free when used with web search or web fetch. When web_search_20260209 (or later) or web_fetch_20260209 (or later) is included in your API request, there are no additional charges for code execution tool calls beyond the standard input and output token costs.@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN When used without these tools, code execution is billed by execution time, tracked separately from token usage:@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Execution time has a minimum of 5 minutes@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Each organization receives 1,550 free hours of usage per month@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Additional usage beyond 1,550 hours is billed at $0.05 per hour, per container@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * If files are included in the request, execution time is billed even if the tool is not invoked, due to files being preloaded onto the container@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Code execution usage is tracked in the response:@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN { "usage" : { "input_tokens" : 105 , "output_tokens" : 239 , "server_tool_use" : { "code_execution_requests" : 1 } } }@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN \ue056@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Tokens@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN All tokens consumed by a Claude Managed Agents session are billed at the rates shown in Model pricing. Prompt caching multipliers apply identically. Web search triggered inside a session incurs the standard $10 per 1,000 searches. On Claude Platform on AWS, session token and runtime charges convert to Claude Consumption Units at the standard rate.@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN The following Messages API modifiers do not apply to Claude Managed Agents sessions:@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Modifier Why it doesn't apply@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Batch API discount Sessions are stateful and interactive. There is no batch mode.@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Fast mode premium Inference speed is managed by the runtime.@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Data residency multiplier inference_geo is a Messages API request field.@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Cloud platform pricing Not available on partner-operated cloud platforms.@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Text editor tool@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN The text editor tool uses the same pricing structure as other tools used with Claude. It follows the standard input and output token pricing based on the Claude model you're using.@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN In addition to the base tokens, the following additional input tokens are needed for the text editor tool:@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Tool Additional input tokens@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN text_editor_20250429 (Claude 4.x) 700 tokens@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN See tool use pricing for complete pricing details.@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Session runtime@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN SKU Rate Metering@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Session runtime $0.08 per session-hour running status duration@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Runtime is measured to the millisecond and accrues only while the session's status is running. Time spent idle (waiting for your next message or a tool confirmation), rescheduling, or terminated does not count toward runtime.@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Web search tool@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Web search usage is charged in addition to token usage:@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN { "usage" : { "input_tokens" : 105 , "output_tokens" : 6039 , "cache_read_input_tokens" : 7123 , "cache_creation_input_tokens" : 7345 , "server_tool_use" : { "web_search_requests" : 1 } } }@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN \ue056@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Web search is available on the Claude API for $10 per 1,000 searches, plus standard token costs for search-generated content. Web search results retrieved throughout a conversation are counted as input tokens, in search iterations executed during a single turn and in subsequent conversation turns.@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Each web search counts as one use, regardless of the number of results returned. If an error occurs during web search, the web search will not be billed.@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN \ue09a@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Web fetch tool@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Web fetch usage has no additional charges beyond standard token costs:@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN { "usage" : { "input_tokens" : 25039 , "output_tokens" : 931 , "cache_read_input_tokens" : 0 , "cache_creation_input_tokens" : 0 , "server_tool_use" : { "web_fetch_requests" : 1 } } }@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN \ue056@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN The web fetch tool is available on the Claude API at no additional cost. You only pay standard token costs for the fetched content that becomes part of your conversation context.@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN To protect against inadvertently fetching large content that would consume excessive tokens, use the max_content_tokens parameter to set appropriate limits based on your use case and budget considerations.@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Example token usage for typical content:@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Average web page (10 kB): ~2,500 tokens@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Large documentation page (100 kB): ~25,000 tokens@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Research paper PDF (500 kB): ~125,000 tokens@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN \ue09a@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Computer use tool@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Computer use follows the standard tool use pricing. When using the computer use tool:@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN System prompt overhead: The computer use beta adds 466-499 tokens to the system prompt@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Computer use tool token usage:@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Model Input tokens per tool definition@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Claude 4.x models 735 tokens@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Additional token consumption:@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Screenshot images (see Vision pricing)@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Tool execution results returned to Claude@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Session runtime replaces the Code Execution container-hour billing model when using Claude Managed Agents. You are not separately billed for container hours on top of session runtime.@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN If you're also using bash or text editor tools alongside computer use, those tools have their own token costs as documented in their respective pages.@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Worked example@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN A one-hour coding session using Claude Opus 4.8 that consumes 50,000 input tokens and 15,000 output tokens:@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Line item Calculation Cost@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Input tokens 50,000 \u00d7 $5 / 1,000,000 $0.25@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Output tokens 15,000 \u00d7 $25 / 1,000,000 $0.375@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Session runtime 1.0 hour \u00d7 $0.08 $0.08@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Total $0.705@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN If prompt caching is active and 40,000 of the input tokens are cache reads:@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Line item Calculation Cost@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Uncached input tokens 10,000 \u00d7 $5 / 1,000,000 $0.05@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Cache read tokens 40,000 \u00d7 $5 \u00d7 0.1 / 1,000,000 $0.02@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Output tokens 15,000 \u00d7 $25 / 1,000,000 $0.375@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Session runtime 1.0 hour \u00d7 $0.08 $0.08@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Total $0.525@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Claude Managed Agents pricing@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Claude Managed Agents is billed on two dimensions: tokens and session runtime.@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN \ue09a@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Tokens@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN All tokens consumed by a Claude Managed Agents session are billed at the rates shown in Model pricing. Prompt caching multipliers apply identically. Web search triggered inside a session incurs the standard $10 per 1,000 searches. On Claude Platform on AWS, session token and runtime charges convert to Claude Consumption Units at the standard rate.@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN The following Messages API modifiers do not apply to Claude Managed Agents sessions:@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Modifier Why it doesn't apply@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Batch API discount Sessions are stateful and interactive. There is no batch mode.@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Fast mode premium Inference speed is managed by the runtime.@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Data residency multiplier inference_geo is a Messages API request field.@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Cloud platform pricing Not available on partner-operated cloud platforms.@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN \ue09a@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Session runtime@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN SKU Rate Metering@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Session runtime $0.08 per session-hour running status duration@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Runtime is measured to the millisecond and accrues only while the session's status is running. Time spent idle (waiting for your next message or a tool confirmation), rescheduling, or terminated does not count toward runtime.@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Example calculation for processing 10,000 support tickets:@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Average ~3,700 tokens per conversation@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Using Claude Haiku 4.5 at $1/MTok input, $5/MTok output@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Total cost: ~$37.00 per 10,000 tickets@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN For a detailed walkthrough of this calculation, see the customer support agent guide.@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Session runtime replaces the Code Execution container-hour billing model when using Claude Managed Agents. You are not separately billed for container hours on top of session runtime.@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Additional pricing considerations@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN \ue09a@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Cost optimization strategies@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN When building agents with Claude:@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN 1. Use appropriate models: Choose Haiku for simple tasks, Sonnet for most production workloads, and Opus for the most complex reasoning@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN 2. Implement prompt caching: Reduce costs for repeated context@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN 3. Batch operations: Use the Batch API for non-time-sensitive tasks@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN 4. Monitor usage patterns: Track token consumption to identify optimization opportunities@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN \ue097@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN For high-volume agent applications, contact the enterprise sales team for custom pricing arrangements.@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN \ue09a@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Rate limits@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Rate limits vary by usage tier and affect how many requests you can make:@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Start tier: Entry-level limits for getting started@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Build tier: Increased limits for growing applications@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Scale tier: Highest standard limits for production workloads@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN For detailed rate limit information, see Rate limits.@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN For limits beyond the Scale tier or custom pricing arrangements, contact the sales team.@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN \ue09a@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Volume discounts@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Volume discounts may be available for high-volume users. These are negotiated on a case-by-case basis.@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Standard usage tiers use the pricing shown in Model pricing@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Enterprise customers can contact sales for custom pricing@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Academic and research discounts may be available@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN \ue09a@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Enterprise pricing@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN For enterprise customers with specific needs:@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Custom rate limits@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Volume discounts@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Dedicated support@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Custom terms@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Contact the sales team at sales@anthropic.com or through the Claude Console to discuss enterprise pricing options.@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN \ue09a@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Billing and payment@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Billing is based on actual monthly usage@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * All payments are in USD@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Credit card and invoicing options available@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Usage tracking available in the Claude Console@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN \ue09a@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Frequently asked questions@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN How is token usage calculated?@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Tokens are pieces of text that models process. As a rough estimate, 1 token is approximately 4 characters or 0.75 words in English. The exact count varies by language and content type.@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Are there free tiers or trials?@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN New users receive a small amount of free credits to test the API. Contact sales for information about extended trials for enterprise evaluation.@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN How do discounts stack?@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Batch API and prompt caching discounts can be combined. For example, using both features together provides significant cost savings compared to standard API calls. See prompt caching pricing for how the multipliers interact.@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN What payment methods are accepted?@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Major credit cards are accepted for standard accounts. Enterprise customers can arrange invoicing and other payment methods.@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN For additional questions about pricing, contact support@anthropic.com.@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Was this page helpful?@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN \ue0fb \ue0f9@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Model pricing@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Cloud platform pricing@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Claude Platform on AWS pricing@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Inference geography@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Private offers@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Claude in Microsoft Foundry pricing@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Inference geography@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Feature-specific pricing@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Prompt caching@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Data residency pricing@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Fast mode pricing@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Batch processing@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Long context pricing@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Tool use pricing@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Specific tool pricing@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Claude Managed Agents pricing@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Tokens@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Session runtime@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Worked example@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Additional pricing considerations@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Cost optimization strategies@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Rate limits@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Volume discounts@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Enterprise pricing@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Billing and payment@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Frequently asked questions@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Claude Platform Docs@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Solutions@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * AI agents@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Code modernization@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Coding@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Customer support@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Education@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Financial services@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Government@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Life sciences@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Partners@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Claude on AWS@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Claude on Google Cloud@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Learn@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Blog@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Courses@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Use cases@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Connectors@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Customer stories@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Engineering at Anthropic@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Events@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Powered by Claude@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Service partners@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Startups program@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Company@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Anthropic@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Careers@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Economic Futures@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Research@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * News@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Responsible Scaling Policy@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Security and compliance@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Transparency@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Learn@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Blog@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Courses@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Use cases@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Connectors@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Customer stories@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Engineering at Anthropic@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Events@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Powered by Claude@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Service partners@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Startups program@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Help and security@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Availability@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Status@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Support@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Discord@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Terms and policies@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Privacy policy@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Responsible disclosure policy@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Terms of service: Commercial@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Terms of service: Consumer@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Usage policy@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Worked example@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN A one-hour coding session using Claude Opus 4.8 that consumes 50,000 input tokens and 15,000 output tokens:@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Line item Calculation Cost@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Input tokens 50,000 \u00d7 $5 / 1,000,000 $0.25@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Output tokens 15,000 \u00d7 $25 / 1,000,000 $0.375@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Session runtime 1.0 hour \u00d7 $0.08 $0.08@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Total $0.705@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN If prompt caching is active and 40,000 of the input tokens are cache reads:@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Line item Calculation Cost@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Uncached input tokens 10,000 \u00d7 $5 / 1,000,000 $0.05@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Cache read tokens 40,000 \u00d7 $5 \u00d7 0.1 / 1,000,000 $0.02@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Output tokens 15,000 \u00d7 $25 / 1,000,000 $0.375@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Session runtime 1.0 hour \u00d7 $0.08 $0.08@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Total $0.525@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN \ue08f@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Example calculation for processing 10,000 support tickets:@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Average ~3,700 tokens per conversation@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Using Claude Haiku 4.5 at $1/MTok input, $5/MTok output@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Total cost: ~$37.00 per 10,000 tickets@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN For a detailed walkthrough of this calculation, see the customer support agent guide.@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN \ue09a@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Additional pricing considerations@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN \ue09a@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Cost optimization strategies@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN When building agents with Claude:@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN 1. Use appropriate models: Choose Haiku for simple tasks, Sonnet for most production workloads, and Opus for the most complex reasoning@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN 2. Implement prompt caching: Reduce costs for repeated context@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN 3. Batch operations: Use the Batch API for non-time-sensitive tasks@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN 4. Monitor usage patterns: Track token consumption to identify optimization opportunities@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN \ue097@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN For high-volume agent applications, contact the enterprise sales team for custom pricing arrangements.@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN \ue09a@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Rate limits@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Rate limits vary by usage tier and affect how many requests you can make:@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Start tier: Entry-level limits for getting started@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Build tier: Increased limits for growing applications@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Scale tier: Highest standard limits for production workloads@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN For detailed rate limit information, see Rate limits.@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN For limits beyond the Scale tier or custom pricing arrangements, contact the sales team.@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN \ue09a@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Volume discounts@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Volume discounts may be available for high-volume users. These are negotiated on a case-by-case basis.@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Standard usage tiers use the pricing shown in Model pricing@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Enterprise customers can contact sales for custom pricing@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Academic and research discounts may be available@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN \ue09a@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Enterprise pricing@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN For enterprise customers with specific needs:@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Custom rate limits@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Volume discounts@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Dedicated support@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Custom terms@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Contact the sales team at sales@anthropic.com or through the Claude Console to discuss enterprise pricing options.@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN \ue09a@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Billing and payment@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Billing is based on actual monthly usage@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * All payments are in USD@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Credit card and invoicing options available@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Usage tracking available in the Claude Console@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN \ue09a@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Frequently asked questions@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN How is token usage calculated?@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Tokens are pieces of text that models process. As a rough estimate, 1 token is approximately 4 characters or 0.75 words in English. The exact count varies by language and content type.@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Are there free tiers or trials?@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN New users receive a small amount of free credits to test the API. Contact sales for information about extended trials for enterprise evaluation.@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN How do discounts stack?@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Batch API and prompt caching discounts can be combined. For example, using both features together provides significant cost savings compared to standard API calls. See prompt caching pricing for how the multipliers interact.@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN What payment methods are accepted?@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Major credit cards are accepted for standard accounts. Enterprise customers can arrange invoicing and other payment methods.@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN For additional questions about pricing, contact support@anthropic.com.@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Was this page helpful?@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN \ue0fb \ue0f9@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Model pricing@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Cloud platform pricing@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Claude Platform on AWS pricing@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Inference geography@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Private offers@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Claude in Microsoft Foundry pricing@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Inference geography@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Feature-specific pricing@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Prompt caching@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Data residency pricing@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Fast mode pricing@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Batch processing@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Long context pricing@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Tool use pricing@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Specific tool pricing@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Claude Managed Agents pricing@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Tokens@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Session runtime@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Worked example@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Additional pricing considerations@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Cost optimization strategies@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Rate limits@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Volume discounts@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Enterprise pricing@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Billing and payment@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Frequently asked questions@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Claude Platform Docs@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Solutions@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * AI agents@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Code modernization@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Coding@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Customer support@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Education@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Financial services@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Government@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Life sciences@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Partners@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Claude on AWS@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Claude on Google Cloud@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Learn@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Blog@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Courses@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Use cases@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Connectors@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Customer stories@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Engineering at Anthropic@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Events@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Powered by Claude@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Service partners@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Startups program@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Company@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Anthropic@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Careers@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Economic Futures@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Research@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * News@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Responsible Scaling Policy@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Security and compliance@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Transparency@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Learn@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Blog@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Courses@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Use cases@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Connectors@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Customer stories@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Engineering at Anthropic@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Events@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Powered by Claude@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Service partners@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Startups program@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Help and security@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Availability@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Status@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Support@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Discord@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Terms and policies@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Privacy policy@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Responsible disclosure policy@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Terms of service: Commercial@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Terms of service: Consumer@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Usage policy@changed_PLACEMARKER_CLOSED\n@changed_into_PLACEMARKER_OPEN * If files are included in the request, execution time is billed even if the tool is not invoked, due tobecause files beingare preloaded onto the container@changed_into_PLACEMARKER_CLOSED", "markup_text_links_to_html_links": false, "restock": {}, "notification_timestamp": 1784178775.562368, "prev_snapshot": " Cookie settings\n\n We use cookies to deliver and improve our services, analyze site usage, and if you agree, to customize or personalize your experience and market our services to you. You can read our Cookie Policy here.\n\n Customize Customize Cookie Settings Reject Reject All Cookies Accept Accept All Cookies\n Claude Platform Docs\n * Messages\n * Managed Agents\n * Admin\n * Resources \ue027\n \ue04a API reference\n \ue053 English \ue027 \ue04f ConsoleLog in\n \ue0d3 \ue062\n \ue0d3\n Search...\n \u2318K\n Models\n Models overview Model IDs and versioning Choosing a model Introducing Claude Fable 5 and Claude Mythos 5 What's new in Claude Opus 4.8 What's new in Claude Sonnet 5 Upgrade between model versions Model deprecations Model cards System prompts Pricing\n \ue104\n Log in\n \ue0dd Models & pricing \ue027 Pricing\n Models & pricing / Models\n\n Pricing\n\n Cookie settings\n\n We use cookies to deliver and improve our services, analyze site usage, and if you agree, to customize or personalize your experience and market our services to you. You can read our Cookie Policy here.\n\n Customize Customize Cookie Settings Reject Reject All Cookies Accept Accept All Cookies\n Claude Platform Docs\n * Messages\n * Managed Agents\n * Admin\n * Resources \ue027\n \ue04a API reference\n \ue053 English \ue027 \ue04f ConsoleLog in\n \ue0d3 \ue062\n \ue0d3\n Search...\n \u2318K\n Models\n Models overview Model IDs and versioning Choosing a model Introducing Claude Fable 5 and Claude Mythos 5 What's new in Claude Opus 4.8 What's new in Claude Sonnet 5 Upgrade between model versions Model deprecations Model cards System prompts Pricing\n \ue104\n Log in\n \ue0dd Models & pricing \ue027 Pricing\n Models & pricing / Models\n\n Pricing\n\n Copy page\n \ue027\n Learn about Anthropic's pricing structure for models and features\n Learn about Anthropic's pricing structure for models and features\n Copy page\n \ue027\n\n This page provides detailed pricing information for Anthropic's models and features. All prices are in USD.\n\n For the most current pricing information, visit claude.com/pricing.\n\n \ue09a\n Model pricing\n\n The following table shows pricing for all Claude models:\n\n Model Base Input Tokens 5m Cache Writes 1h Cache Writes Cache Hits & Refreshes Output Tokens\n Claude Fable 5 $10 / MTok $12.50 / MTok $20 / MTok $1 / MTok $50 / MTok\n Claude Mythos 5 (limited availability) $10 / MTok $12.50 / MTok $20 / MTok $1 / MTok $50 / MTok\n Claude Opus 4.8 $5 / MTok $6.25 / MTok $10 / MTok $0.50 / MTok $25 / MTok\n Claude Opus 4.7 $5 / MTok $6.25 / MTok $10 / MTok $0.50 / MTok $25 / MTok\n Claude Opus 4.6 $5 / MTok $6.25 / MTok $10 / MTok $0.50 / MTok $25 / MTok\n Claude Opus 4.5 $5 / MTok $6.25 / MTok $10 / MTok $0.50 / MTok $25 / MTok\n Claude Opus 4.1 (deprecated) $15 / MTok $18.75 / MTok $30 / MTok $1.50 / MTok $75 / MTok\n Claude Opus 4 (retired, except on Google Cloud) $15 / MTok $18.75 / MTok $30 / MTok $1.50 / MTok $75 / MTok\n Claude Sonnet 5 $2 / MTok $2.50 / MTok $4 / MTok $0.20 / MTok $10 / MTok\n through August 31, 2026\n Claude Sonnet 5 $3 / MTok $3.75 / MTok $6 / MTok $0.30 / MTok $15 / MTok\n starting September 1, 2026\n Claude Sonnet 4.6 $3 / MTok $3.75 / MTok $6 / MTok $0.30 / MTok $15 / MTok\n Claude Sonnet 4.5 $3 / MTok $3.75 / MTok $6 / MTok $0.30 / MTok $15 / MTok\n Claude Sonnet 4 (retired, except on Bedrock and Google Cloud) $3 / MTok $3.75 / MTok $6 / MTok $0.30 / MTok $15 / MTok\n Claude Haiku 4.5 $1 / MTok $1.25 / MTok $2 / MTok $0.10 / MTok $5 / MTok\n Claude Haiku 3.5 (retired, except on Bedrock and Google Cloud) $0.80 / MTok $1 / MTok $1.60 / MTok $0.08 / MTok $4 / MTok\n\n \ue08f\n\n Introductory pricing of $2/$10 per million input/output tokens is in effect through August 31, 2026, after which the standard pricing of $3/$15 per million input/output tokens will take effect.\n\n \ue08f\n\n MTok = Million tokens. The "Base Input Tokens" column shows standard input pricing, the "5m Cache Writes", "1h Cache Writes", and "Cache Hits & Refreshes" columns are specific to prompt caching, and "Output Tokens" shows output pricing. See prompt caching pricing for an explanation of the cache columns and pricing multipliers.\n\n \ue08f\n\n Claude Opus 4.7 and later Opus models, Claude Fable 5, Claude Mythos 5, Claude Mythos Preview, and Claude Sonnet 5 use a newer tokenizer that contributes to their improved performance on a wide range of tasks. This tokenizer produces approximately 30% more tokens for the same text. The exact increase depends on the content and workload shape. Claude Sonnet 4.6 and earlier models use the previous tokenizer.\n\n For Claude Platform on AWS pricing, see Claude Platform on AWS pricing.\n\n This page provides detailed pricing information for Anthropic's models and features. All prices are in USD.\n\n For the most current pricing information, visit claude.com/pricing.\n Cloud platform pricing\n\n This section covers partner-operated cloud platforms, where the cloud provider invoices you. For Anthropic-operated cloud platforms billed through a marketplace, see Claude Platform on AWS pricing and Claude in Microsoft Foundry pricing.\n\n Claude models are available on Amazon Bedrock and Google Cloud. For official pricing, visit:\n\n * Amazon Bedrock pricing\n * Google Cloud pricing\n \ue08f\n\n Regional and multi-region endpoint pricing for Claude 4.5 models and beyond\n\n Starting with Claude Sonnet 4.5, Haiku 4.5, and Opus 4.5:\n\n * Bedrock offers two endpoint types: global endpoints (dynamic routing for maximum availability) and regional endpoints (guaranteed data routing through specific geographic regions).\n * Google Cloud offers three endpoint types: global endpoints, multi-region endpoints (dynamic routing within a geographic area), and regional endpoints.\n\n Regional and multi-region endpoints include a 10% premium over global endpoints. The Claude API (first-party) is global by default; for first-party data residency options and pricing, see Data residency pricing.\n\n Scope: This pricing structure applies to Claude Sonnet 4.5, Haiku 4.5, Opus 4.5, and all future models. Earlier models (Claude Opus 4.1 (deprecated) and prior releases) retain their existing pricing.\n\n For implementation details and code examples:\n\n * Amazon Bedrock global vs regional endpoints for Opus 4.7, Haiku 4.5, and later models, or the legacy integration for all other models on Bedrock\n * Google Cloud global, multi-region, and regional endpoints\n\n \ue09a\n Claude Platform on AWS pricing\n\n Claude Platform on AWS bills through AWS Marketplace using Claude Consumption Units (CCUs). Anthropic rates your token usage in USD at standard per-model, per-feature rates, applies any negotiated discount, converts the result to CCUs at $0.01 per CCU, and reports the CCU quantity to AWS Marketplace hourly. Your AWS bill shows a single CCU line item.\n\n Concept Details\n Billing unit Claude Consumption Unit (CCU)\n CCU price $0.01 per CCU (fixed; discounts apply at token-to-CCU conversion, not to the CCU price)\n Conversion Token usage rated in USD at standard per-model, per-feature rates (same as Claude API pricing), then converted to CCUs at $0.01 per CCU\n Billing cadence Hourly metering to AWS Marketplace; monthly invoices\n Payment model Arrears only (postpaid); no prepaid credits\n Discounts Applied as fewer CCUs metered\n Tax Pre-tax metering; AWS Marketplace handles tax\n Cost visibility Real-time breakdown in the Claude Console (access through the AWS Console); AWS Cost Explorer shows aggregated CCU\n Model pricing\n\n The following table shows pricing for all Claude models:\n\n Model Base Input Tokens 5m Cache Writes 1h Cache Writes Cache Hits & Refreshes Output Tokens\n Claude Fable 5 $10 / MTok $12.50 / MTok $20 / MTok $1 / MTok $50 / MTok\n Claude Mythos 5 (limited availability) $10 / MTok $12.50 / MTok $20 / MTok $1 / MTok $50 / MTok\n Claude Opus 4.8 $5 / MTok $6.25 / MTok $10 / MTok $0.50 / MTok $25 / MTok\n Claude Opus 4.7 $5 / MTok $6.25 / MTok $10 / MTok $0.50 / MTok $25 / MTok\n Claude Opus 4.6 $5 / MTok $6.25 / MTok $10 / MTok $0.50 / MTok $25 / MTok\n Claude Opus 4.5 $5 / MTok $6.25 / MTok $10 / MTok $0.50 / MTok $25 / MTok\n Claude Opus 4.1 (deprecated) $15 / MTok $18.75 / MTok $30 / MTok $1.50 / MTok $75 / MTok\n Claude Opus 4 (retired, except on Google Cloud) $15 / MTok $18.75 / MTok $30 / MTok $1.50 / MTok $75 / MTok\n Claude Sonnet 5 $2 / MTok $2.50 / MTok $4 / MTok $0.20 / MTok $10 / MTok\n through August 31, 2026\n Claude Sonnet 5 $3 / MTok $3.75 / MTok $6 / MTok $0.30 / MTok $15 / MTok\n starting September 1, 2026\n Claude Sonnet 4.6 $3 / MTok $3.75 / MTok $6 / MTok $0.30 / MTok $15 / MTok\n Claude Sonnet 4.5 $3 / MTok $3.75 / MTok $6 / MTok $0.30 / MTok $15 / MTok\n Claude Sonnet 4 (retired, except on Bedrock and Google Cloud) $3 / MTok $3.75 / MTok $6 / MTok $0.30 / MTok $15 / MTok\n Claude Haiku 4.5 $1 / MTok $1.25 / MTok $2 / MTok $0.10 / MTok $5 / MTok\n Claude Haiku 3.5 (retired, except on Bedrock and Google Cloud) $0.80 / MTok $1 / MTok $1.60 / MTok $0.08 / MTok $4 / MTok\n Claude Consumption Units. If Customer accesses the Services through certain Marketplace Platforms (e.g., Claude Platform on AWS), usage will be invoiced in Claude Consumption Units ("CCU") rather than per MTok. A CCU is a unit of measure used solely for Marketplace Platform invoicing. One hundred (100) CCU represents $1.00 USD of fees owed for the Services, calculated at the applicable prices on claude.com/pricing#api, after application of any discounts.\n Introductory pricing of $2/$10 per million input/output tokens is in effect through August 31, 2026, after which the standard pricing of $3/$15 per million input/output tokens will take effect.\n\n \ue08f\n\n MTok = Million tokens. The "Base Input Tokens" column shows standard input pricing, the "5m Cache Writes", "1h Cache Writes", and "Cache Hits & Refreshes" columns are specific to prompt caching, and "Output Tokens" shows output pricing. See prompt caching pricing for an explanation of the cache columns and pricing multipliers.\n\n \ue08f\n\n Claude Opus 4.7 and later Opus models, Claude Fable 5, Claude Mythos 5, Claude Mythos Preview, and Claude Sonnet 5 use a newer tokenizer that contributes to their improved performance on a wide range of tasks. This tokenizer produces approximately 30% more tokens for the same text. The exact increase depends on the content and workload shape. Claude Sonnet 4.6 and earlier models use the previous tokenizer.\n\n For Claude Platform on AWS pricing, see Claude Platform on AWS pricing.\n Inference geography\n\n For Claude Opus 4.6, Claude Sonnet 4.6, and later models, using inference_geo: "us" applies a 1.1x pricing multiplier. inference_geo: "global" (default) uses standard pricing. See Data residency for details.\n Cloud platform pricing\n\n This section covers partner-operated cloud platforms, where the cloud provider invoices you. For Anthropic-operated cloud platforms billed through a marketplace, see Claude Platform on AWS pricing and Claude in Microsoft Foundry pricing.\n\n Claude models are available on Amazon Bedrock and Google Cloud. For official pricing, visit:\n\n * Amazon Bedrock pricing\n * Google Cloud pricing\n \ue08f\n\n Regional and multi-region endpoint pricing for Claude 4.5 models and beyond\n\n Starting with Claude Sonnet 4.5, Haiku 4.5, and Opus 4.5:\n\n * Bedrock offers two endpoint types: global endpoints (dynamic routing for maximum availability) and regional endpoints (guaranteed data routing through specific geographic regions).\n * Google Cloud offers three endpoint types: global endpoints, multi-region endpoints (dynamic routing within a geographic area), and regional endpoints.\n\n Regional and multi-region endpoints include a 10% premium over global endpoints. The Claude API (first-party) is global by default; for first-party data residency options and pricing, see Data residency pricing.\n\n Scope: This pricing structure applies to Claude Sonnet 4.5, Haiku 4.5, Opus 4.5, and all future models. Earlier models (Claude Opus 4.1 (deprecated) and prior releases) retain their existing pricing.\n\n For implementation details and code examples:\n\n * Amazon Bedrock global vs regional endpoints for Opus 4.7, Haiku 4.5, and later models, or the legacy integration for all other models on Bedrock\n * Google Cloud global, multi-region, and regional endpoints\n Private offers\n\n When you sign up on the AWS Console Claude Platform on AWS service page, the AWS Console looks up any private offer associated with your account and prompts you to accept it in AWS Marketplace. Contact your Anthropic account representative for private offer terms.\n\n \ue08f\n\n If you have an existing Amazon Bedrock private offer, contact your Anthropic or AWS account representative before getting started with Claude Platform on AWS to ensure your discounts are applied correctly. Discounts cannot be applied retroactively to usage incurred before your private offer is accepted.\n\n \ue09a\n Claude in Microsoft Foundry pricing\n\n Claude in Microsoft Foundry bills through the Azure Marketplace using Claude Consumption Units (CCUs). Anthropic rates your token usage in USD at standard per-model, per-feature rates, applies any negotiated discount, converts the result to CCUs at $0.01 per CCU, and reports the CCU quantity to the Azure Marketplace hourly. Your Azure bill shows a single CCU line item.\n Claude Platform on AWS pricing\n\n Claude Platform on AWS bills through AWS Marketplace using Claude Consumption Units (CCUs). Anthropic rates your token usage in USD at standard per-model, per-feature rates, applies any negotiated discount, converts the result to CCUs at $0.01 per CCU, and reports the CCU quantity to AWS Marketplace hourly. Your AWS bill shows a single CCU line item.\n Billing cadence Hourly metering to the AzureAWS Marketplace; monthly invoices\n Tax Pre-tax metering; Azure Marketplace handles tax\n Cost visibility Azure Cost Management shows aggregated CCU\n Tax Pre-tax metering; AWS Marketplace handles tax\n Cost visibility Real-time breakdown in the Claude Console (access through the AWS Console); AWS Cost Explorer shows aggregated CCU\n Claude Consumption Units. If Customer accesses the Services through certain Marketplace Platforms (e.g., Claude Platform on AWS, Claude in Microsoft Foundry),AWS), usage will be invoiced in Claude Consumption Units ("CCU") rather than per MTok. A CCU is a unit of measure used solely for Marketplace Platform invoicing. One hundred (100) CCU represents $1.00 USD of fees owed for the Services, calculated at the applicable prices on claude.com/pricing#api, after application of any discounts.\n Deployments hosted on Azure canFor useClaude theOpus US Data4.6, Zone Standard deployment type, which keeps inference within theClaude UnitedSonnet States.4.6, Thisand islater equivalentmodels, tousing inference_geo: "us" onapplies thea Claude1.1x APIpricing andmultiplier. appliesinference_geo: the"global" same(default) 1.1xuses pricingstandard multiplier.pricing. See Data residency for details.\n Feature-specific pricing\n Private offers\n\n When you sign up on the AWS Console Claude Platform on AWS service page, the AWS Console looks up any private offer associated with your account and prompts you to accept it in AWS Marketplace. Contact your Anthropic account representative for private offer terms.\n\n \ue08f\n\n If you have an existing Amazon Bedrock private offer, contact your Anthropic or AWS account representative before getting started with Claude Platform on AWS to ensure your discounts are applied correctly. Discounts cannot be applied retroactively to usage incurred before your private offer is accepted.\n Prompt caching\n\n Prompt caching reduces costs and latency by reusing previously processed portions of your prompt across API calls. Instead of reprocessing the same large system prompt, document, or conversation history on every request, the API reads from cache at a fraction of the standard input price.\n\n There are two ways to enable prompt caching:\n\n * Automatic caching: Add a single cache_control field at the top level of your request. The system automatically manages cache breakpoints as conversations grow. This is the recommended starting point for most use cases.\n * Explicit cache breakpoints: Place cache_control directly on individual content blocks for fine-grained control over exactly what gets cached.\n\n Prompt caching uses the following pricing multipliers relative to base input token rates:\n\n Cache operation Multiplier Duration\n 5-minute cache write 1.25x base input price Cache valid for 5 minutes\n 1-hour cache write 2x base input price Cache valid for 1 hour\n Cache read (hit) 0.1x base input price Same duration as the preceding write\n\n\n Cache write tokens are charged when content is first stored. Cache read tokens are charged when a subsequent request retrieves the cached content. A cache hit costs 10% of the standard input price, which means caching pays off after just one cache read for the 5-minute duration (1.25x write), or after two cache reads for the 1-hour duration (2x write).\n\n These multipliers stack with other pricing modifiers, including the Batch API discount and data residency.\n\n For implementation details, supported models, and code examples, see Prompt caching.\n Claude in Microsoft Foundry pricing\n\n Claude in Microsoft Foundry bills through the Azure Marketplace using Claude Consumption Units (CCUs). Anthropic rates your token usage in USD at standard per-model, per-feature rates, applies any negotiated discount, converts the result to CCUs at $0.01 per CCU, and reports the CCU quantity to the Azure Marketplace hourly. Your Azure bill shows a single CCU line item.\n\n Concept Details\n Billing unit Claude Consumption Unit (CCU)\n CCU price $0.01 per CCU (fixed; discounts apply at token-to-CCU conversion, not to the CCU price)\n Conversion Token usage rated in USD at standard per-model, per-feature rates (same as Claude API pricing), then converted to CCUs at $0.01 per CCU\n Billing cadence Hourly metering to the Azure Marketplace; monthly invoices\n Payment model Arrears only (postpaid); no prepaid credits\n Discounts Applied as fewer CCUs metered\n Tax Pre-tax metering; Azure Marketplace handles tax\n Cost visibility Azure Cost Management shows aggregated CCU\n\n \ue08f\n\n Claude Consumption Units. If Customer accesses the Services through certain Marketplace Platforms (e.g., Claude Platform on AWS, Claude in Microsoft Foundry), usage will be invoiced in Claude Consumption Units ("CCU") rather than per MTok. A CCU is a unit of measure used solely for Marketplace Platform invoicing. One hundred (100) CCU represents $1.00 USD of fees owed for the Services, calculated at the applicable prices on claude.com/pricing#api, after application of any discounts.\n Data residency pricing\n\n For Claude Opus 4.6, Claude Sonnet 4.6, and later models, specifying US-only inference through the inference_geo parameter incurs a 1.1x multiplier on all token pricing categories, including input tokens, output tokens, cache writes, and cache reads. Global routing (the default) uses standard pricing.\n\n This applies to the Claude API (first-party) and Claude Platform on AWS. On Claude in Microsoft Foundry, the same 1.1x multiplier applies to deployments that use the US Data Zone Standard deployment type (see Inference geography). Partner-operated platforms (Bedrock and Google Cloud) have independent regional pricing. See Bedrock and Google Cloud for details. Earlier models do not support the inference_geo parameter and always use standard pricing; requests that include the parameter on these models return a 400 error.\n\n For more information, see Data residency.\n Inference geography\n\n Deployments hosted on Azure can use the US Data Zone Standard deployment type, which keeps inference within the United States. This is equivalent to inference_geo: "us" on the Claude API and applies the same 1.1x pricing multiplier. See Data residency for details.\n Fast mode pricing\n\n Fast mode, in research preview, provides significantly faster output for Claude Opus 4.8 and Claude Opus 4.7 at premium pricing. Fast mode pricing applies across the full context window, including requests over 200k input tokens. Fast mode is not available on Claude Platform on AWS.\n\n Model Input Output\n Claude Opus 4.8 $10 / MTok $50 / MTok\n Claude Opus 4.7 $30 / MTok $150 / MTok\n\n\n Fast mode for Claude Opus 4.7 is deprecated and will be removed on July 24, 2026. As of June 29, 2026, fast mode is not available on Claude Opus 4.6: requests to claude-opus-4-6 with speed: "fast" run at standard speed and are billed at standard rates. See Fast mode.\n\n Fast mode pricing stacks with other pricing modifiers:\n\n * Prompt caching multipliers apply on top of fast mode pricing\n * Data residency multipliers apply on top of fast mode pricing\n\n Fast mode is not available with the Batch API.\n\n For more information, see Fast mode.\n Feature-specific pricing\n\n \ue09a\n Prompt caching\n\n Prompt caching reduces costs and latency by reusing previously processed portions of your prompt across API calls. Instead of reprocessing the same large system prompt, document, or conversation history on every request, the API reads from cache at a fraction of the standard input price.\n\n There are two ways to enable prompt caching:\n\n * Automatic caching: Add a single cache_control field at the top level of your request. The system automatically manages cache breakpoints as conversations grow. This is the recommended starting point for most use cases.\n * Explicit cache breakpoints: Place cache_control directly on individual content blocks for fine-grained control over exactly what gets cached.\n\n Prompt caching uses the following pricing multipliers relative to base input token rates:\n\n Cache operation Multiplier Duration\n 5-minute cache write 1.25x base input price Cache valid for 5 minutes\n 1-hour cache write 2x base input price Cache valid for 1 hour\n Cache read (hit) 0.1x base input price Same duration as the preceding write\n\n\n Cache write tokens are charged when content is first stored. Cache read tokens are charged when a subsequent request retrieves the cached content. A cache hit costs 10% of the standard input price, which means caching pays off after just one cache read for the 5-minute duration (1.25x write), or after two cache reads for the 1-hour duration (2x write).\n\n These multipliers stack with other pricing modifiers, including the Batch API discount and data residency.\n\n For implementation details, supported models, and code examples, see Prompt caching.\n Batch processing\n\n The Batch API allows asynchronous processing of large volumes of requests with a 50% discount on both input and output tokens.\n\n Model Batch input Batch output\n Claude Fable 5 $5 / MTok $25 / MTok\n Claude Mythos 5 (limited availability) $5 / MTok $25 / MTok\n Claude Opus 4.8 $2.50 / MTok $12.50 / MTok\n Claude Opus 4.7 $2.50 / MTok $12.50 / MTok\n Claude Opus 4.6 $2.50 / MTok $12.50 / MTok\n Claude Opus 4.5 $2.50 / MTok $12.50 / MTok\n Claude Opus 4.1 (deprecated) $7.50 / MTok $37.50 / MTok\n Claude Opus 4 (retired, except on Google Cloud) $7.50 / MTok $37.50 / MTok\n Claude Sonnet 5 $1 / MTok $5 / MTok\n through August 31, 2026\n Claude Sonnet 5 $1.50 / MTok $7.50 / MTok\n starting September 1, 2026\n Claude Sonnet 4.6 $1.50 / MTok $7.50 / MTok\n Claude Sonnet 4.5 $1.50 / MTok $7.50 / MTok\n Claude Sonnet 4 (retired, except on Bedrock and Google Cloud) $1.50 / MTok $7.50 / MTok\n Claude Haiku 4.5 $0.50 / MTok $2.50 / MTok\n Claude Haiku 3.5 (retired, except on Bedrock and Google Cloud) $0.40 / MTok $2 / MTok\n\n\n For more information about batch processing, see Batch processing.\n Data residency pricing\n\n For Claude Opus 4.6, Claude Sonnet 4.6, and later models, specifying US-only inference through the inference_geo parameter incurs a 1.1x multiplier on all token pricing categories, including input tokens, output tokens, cache writes, and cache reads. Global routing (the default) uses standard pricing.\n\n This applies to the Claude API (first-party) and Claude Platform on AWS. On Claude in Microsoft Foundry, the same 1.1x multiplier applies to deployments that use the US Data Zone Standard deployment type (see Inference geography). Partner-operated platforms (Bedrock and Google Cloud) have independent regional pricing. See Bedrock and Google Cloud for details. Earlier models do not support the inference_geo parameter and always use standard pricing; requests that include the parameter on these models return a 400 error.\n\n For more information, see Data residency.\n\n \ue09a\n Fast mode pricing\n\n Fast mode, in research preview, provides significantly faster output for Claude Opus 4.8 and Claude Opus 4.7 at premium pricing. Fast mode pricing applies across the full context window, including requests over 200k input tokens. Fast mode is not available on Claude Platform on AWS.\n\n Model Input Output\n Claude Opus 4.8 $10 / MTok $50 / MTok\n Claude Opus 4.7 $30 / MTok $150 / MTok\n\n\n Fast mode for Claude Opus 4.7 is deprecated and will be removed on July 24, 2026. As of June 29, 2026, fast mode is not available on Claude Opus 4.6: requests to claude-opus-4-6 with speed: "fast" run at standard speed and are billed at standard rates. See Fast mode.\n\n Fast mode pricing stacks with other pricing modifiers:\n\n * Prompt caching multipliers apply on top of fast mode pricing\n * Data residency multipliers apply on top of fast mode pricing\n\n Fast mode is not available with the Batch API.\n\n For more information, see Fast mode.\n Long context pricing\n\n Claude Fable 5, Claude Mythos 5, Claude Mythos Preview, Claude Opus 4.8, Opus 4.7, Opus 4.6, Sonnet 5, and Sonnet 4.6 include the full 1M token context window at standard pricing. (A 900k-token request is billed at the same per-token rate as a 9k-token request.) Prompt caching and batch processing discounts apply at standard rates across the full context window.\n\n \ue09a\n Tool use pricing\n\n Tool use requests are priced based on:\n\n 1. The total number of input tokens sent to the model (including in the tools parameter)\n 2. The number of output tokens generated\n 3. For server-side tools, additional usage-based pricing (e.g., web search charges per search performed)\n\n Client-side tools are priced the same as any other Claude API request, while server-side tools may incur additional charges based on their specific usage.\n\n The additional tokens from tool use come from:\n\n * The tools parameter in API requests (tool names, descriptions, and schemas)\n * tool_use content blocks in API requests and responses\n * tool_result content blocks in API requests\n\n When you use tools, the API also automatically includes a special system prompt for the model which enables tool use. The number of tool use tokens required for each model are listed below (excluding the additional tokens listed above). Note that the table assumes at least 1 tool is provided. If no tools are provided, then a tool choice of none uses 0 additional system prompt tokens.\n\n Model Tool choice Tool use system prompt token count\n Claude Opus 4.8 auto, noneany, tool 290 tokens410 tokens\n Claude Opus 4.7 auto, noneany, tool 675 tokens804 tokens\n Claude Opus 4.6 auto, noneany, tool 497 tokens589 tokens\n Claude Opus 4.5 auto, noneany, tool 496 tokens588 tokens\n Claude Opus 4.1 (deprecated) auto, noneany, tool 313 tokens315 tokens\n Claude Opus 4 (retired, except on Google Cloud) auto, noneany, tool 313 tokens315 tokens\n Claude Sonnet 5 auto, noneany, tool 354 tokens474 tokens\n Claude Sonnet 4.6 auto, noneany, tool 497 tokens589 tokens\n Claude Sonnet 4.5 auto, noneany, tool 496 tokens588 tokens\n Claude Sonnet 4 (retired, except on Bedrock and Google Cloud) auto, noneany, tool 313 tokens315 tokens\n Claude Haiku 4.5 auto, noneany, tool 496 tokens588 tokens\n Claude Haiku 3.5 (retired, except on Bedrock and Google Cloud) auto, noneany, tool 264 tokens355 tokens\n\n\n These token counts are added to your normal input and output tokens to calculate the total cost of a request.\n\n For current per-model prices, refer to the model pricing section.\n\n For more information about tool use implementation and best practices, see Tool use.\n Batch processing\n\n The Batch API allows asynchronous processing of large volumes of requests with a 50% discount on both input and output tokens.\n\n Model Batch input Batch output\n Claude Fable 5 $5 / MTok $25 / MTok\n Claude Mythos 5 (limited availability) $5 / MTok $25 / MTok\n Claude Opus 4.8 $2.50 / MTok $12.50 / MTok\n Claude Opus 4.7 $2.50 / MTok $12.50 / MTok\n Claude Opus 4.6 $2.50 / MTok $12.50 / MTok\n Claude Opus 4.5 $2.50 / MTok $12.50 / MTok\n Claude Opus 4.1 (deprecated) $7.50 / MTok $37.50 / MTok\n Claude Opus 4 (retired, except on Google Cloud) $7.50 / MTok $37.50 / MTok\n Claude Sonnet 5 $1 / MTok $5 / MTok\n through August 31, 2026\n Claude Sonnet 5 $1.50 / MTok $7.50 / MTok\n starting September 1, 2026\n Claude Sonnet 4.6 $1.50 / MTok $7.50 / MTok\n Claude Sonnet 4.5 $1.50 / MTok $7.50 / MTok\n Claude Sonnet 4 (retired, except on Bedrock and Google Cloud) $1.50 / MTok $7.50 / MTok\n Claude Haiku 4.5 $0.50 / MTok $2.50 / MTok\n Claude Haiku 3.5 (retired, except on Bedrock and Google Cloud) $0.40 / MTok $2 / MTok\n\n\n For more information about batch processing, see Batch processing.\n Specific tool pricing\n Long context pricing\n\n Claude Fable 5, Claude Mythos 5, Claude Mythos Preview, Claude Opus 4.8, Opus 4.7, Opus 4.6, Sonnet 5, and Sonnet 4.6 include the full 1M token context window at standard pricing. (A 900k-token request is billed at the same per-token rate as a 9k-token request.) Prompt caching and batch processing discounts apply at standard rates across the full context window.\n Bash tool\n\n The bash tool adds 245 input tokens to your API calls.\n\n Additional tokens are consumed by:\n\n * Command outputs (stdout/stderr)\n * Error messages\n * Large file contents\n\n See tool use pricing for complete pricing details.\n\n \ue09a\n Code execution tool\n\n Code execution is free when used with web search or web fetch. When web_search_20260209 (or later) or web_fetch_20260209 (or later) is included in your API request, there are no additional charges for code execution tool calls beyond the standard input and output token costs.\n\n When used without these tools, code execution is billed by execution time, tracked separately from token usage:\n\n * Execution time has a minimum of 5 minutes\n * Each organization receives 1,550 free hours of usage per month\n * Additional usage beyond 1,550 hours is billed at $0.05 per hour, per container\n * If files are included in the request, execution time is billed even if the tool is not invoked, due to files being preloaded onto the container\n\n Code execution usage is tracked in the response:\n\n { "usage" : { "input_tokens" : 105 , "output_tokens" : 239 , "server_tool_use" : { "code_execution_requests" : 1 } } }\n \ue056\n\n \ue09a\n Text editor tool\n\n The text editor tool uses the same pricing structure as other tools used with Claude. It follows the standard input and output token pricing based on the Claude model you're using.\n\n In addition to the base tokens, the following additional input tokens are needed for the text editor tool:\n\n Tool Additional input tokens\n text_editor_20250429 (Claude 4.x) 700 tokens\n\n\n See tool use pricing for complete pricing details.\n Tool use pricing\n\n Tool use requests are priced based on:\n\n 1. The total number of input tokens sent to the model (including in the tools parameter)\n 2. The number of output tokens generated\n 3. For server-side tools, additional usage-based pricing (e.g., web search charges per search performed)\n\n Client-side tools are priced the same as any other Claude API request, while server-side tools may incur additional charges based on their specific usage.\n\n The additional tokens from tool use come from:\n\n * The tools parameter in API requests (tool names, descriptions, and schemas)\n * tool_use content blocks in API requests and responses\n * tool_result content blocks in API requests\n\n When you use tools, the API also automatically includes a special system prompt for the model which enables tool use. The number of tool use tokens required for each model are listed below (excluding the additional tokens listed above). Note that the table assumes at least 1 tool is provided. If no tools are provided, then a tool choice of none uses 0 additional system prompt tokens.\n\n Model Tool choice Tool use system prompt token count\n Claude Opus 4.8 auto, noneany, tool 290 tokens410 tokens\n Claude Opus 4.7 auto, noneany, tool 675 tokens804 tokens\n Claude Opus 4.6 auto, noneany, tool 497 tokens589 tokens\n Claude Opus 4.5 auto, noneany, tool 496 tokens588 tokens\n Claude Opus 4.1 (deprecated) auto, noneany, tool 313 tokens315 tokens\n Claude Opus 4 (retired, except on Google Cloud) auto, noneany, tool 313 tokens315 tokens\n Claude Sonnet 5 auto, noneany, tool 354 tokens474 tokens\n Claude Sonnet 4.6 auto, noneany, tool 497 tokens589 tokens\n Claude Sonnet 4.5 auto, noneany, tool 496 tokens588 tokens\n Claude Sonnet 4 (retired, except on Bedrock and Google Cloud) auto, noneany, tool 313 tokens315 tokens\n Claude Haiku 4.5 auto, noneany, tool 496 tokens588 tokens\n Claude Haiku 3.5 (retired, except on Bedrock and Google Cloud) auto, noneany, tool 264 tokens355 tokens\n\n\n These token counts are added to your normal input and output tokens to calculate the total cost of a request.\n\n For current per-model prices, refer to the model pricing section.\n\n For more information about tool use implementation and best practices, see Tool use.\n Web search tool\n\n Web search usage is charged in addition to token usage:\n\n { "usage" : { "input_tokens" : 105 , "output_tokens" : 6039 , "cache_read_input_tokens" : 7123 , "cache_creation_input_tokens" : 7345 , "server_tool_use" : { "web_search_requests" : 1 } } }\n \ue056\n\n Web search is available on the Claude API for $10 per 1,000 searches, plus standard token costs for search-generated content. Web search results retrieved throughout a conversation are counted as input tokens, in search iterations executed during a single turn and in subsequent conversation turns.\n\n Each web search counts as one use, regardless of the number of results returned. If an error occurs during web search, the web search will not be billed.\n Specific tool pricing\n Web fetch tool\n\n Web fetch usage has no additional charges beyond standard token costs:\n\n { "usage" : { "input_tokens" : 25039 , "output_tokens" : 931 , "cache_read_input_tokens" : 0 , "cache_creation_input_tokens" : 0 , "server_tool_use" : { "web_fetch_requests" : 1 } } }\n \ue056\n\n The web fetch tool is available on the Claude API at no additional cost. You only pay standard token costs for the fetched content that becomes part of your conversation context.\n\n To protect against inadvertently fetching large content that would consume excessive tokens, use the max_content_tokens parameter to set appropriate limits based on your use case and budget considerations.\n\n Example token usage for typical content:\n\n * Average web page (10 kB): ~2,500 tokens\n * Large documentation page (100 kB): ~25,000 tokens\n * Research paper PDF (500 kB): ~125,000 tokens\n\n \ue09a\n Computer use tool\n\n Computer use follows the standard tool use pricing. When using the computer use tool:\n\n System prompt overhead: The computer use beta adds 466-499 tokens to the system prompt\n\n Computer use tool token usage:\n\n Model Input tokens per tool definition\n Claude 4.x models 735 tokens\n\n\n Additional token consumption:\n\n * Screenshot images (see Vision pricing)\n * Tool execution results returned to Claude\n \ue08f\n\n If you're also using bash or text editor tools alongside computer use, those tools have their own token costs as documented in their respective pages.\n Bash tool\n\n The bash tool definition adds the following input tokens to your request. This is in addition to the per-model tool use system prompt that applies whenever any tool is present.\n\n Model Additional input tokens\n Claude Opus 4.7 and Claude Opus 4.8 325 tokens\n Claude Opus 4.6, Claude Sonnet 4.6, and earlier 244 tokens\n\n\n Additional tokens are consumed by:\n\n * Command outputs (stdout/stderr)\n * Error messages\n * Large file contents\n\n See tool use pricing for complete pricing details.\n Claude Managed Agents pricing\n\n Claude Managed Agents is billed on two dimensions: tokens and session runtime.\n Code execution tool\n\n Code execution is free when used with web search or web fetch. When web_search_20260209 (or later) or web_fetch_20260209 (or later) is included in your API request, there are no additional charges for code execution tool calls beyond the standard input and output token costs.\n\n When used without these tools, code execution is billed by execution time, tracked separately from token usage:\n\n * Execution time has a minimum of 5 minutes\n * Each organization receives 1,550 free hours of usage per month\n * Additional usage beyond 1,550 hours is billed at $0.05 per hour, per container\n * If files are included in the request, execution time is billed even if the tool is not invoked, due to files being preloaded onto the container\n\n Code execution usage is tracked in the response:\n\n { "usage" : { "input_tokens" : 105 , "output_tokens" : 239 , "server_tool_use" : { "code_execution_requests" : 1 } } }\n \ue056\n Tokens\n\n All tokens consumed by a Claude Managed Agents session are billed at the rates shown in Model pricing. Prompt caching multipliers apply identically. Web search triggered inside a session incurs the standard $10 per 1,000 searches. On Claude Platform on AWS, session token and runtime charges convert to Claude Consumption Units at the standard rate.\n\n The following Messages API modifiers do not apply to Claude Managed Agents sessions:\n\n Modifier Why it doesn't apply\n Batch API discount Sessions are stateful and interactive. There is no batch mode.\n Fast mode premium Inference speed is managed by the runtime.\n Data residency multiplier inference_geo is a Messages API request field.\n Cloud platform pricing Not available on partner-operated cloud platforms.\n\n Text editor tool\n\n The text editor tool uses the same pricing structure as other tools used with Claude. It follows the standard input and output token pricing based on the Claude model you're using.\n\n In addition to the base tokens, the following additional input tokens are needed for the text editor tool:\n\n Tool Additional input tokens\n text_editor_20250429 (Claude 4.x) 700 tokens\n\n\n See tool use pricing for complete pricing details.\n Session runtime\n\n SKU Rate Metering\n Session runtime $0.08 per session-hour running status duration\n\n\n Runtime is measured to the millisecond and accrues only while the session's status is running. Time spent idle (waiting for your next message or a tool confirmation), rescheduling, or terminated does not count toward runtime.\n\n Web search tool\n\n Web search usage is charged in addition to token usage:\n\n { "usage" : { "input_tokens" : 105 , "output_tokens" : 6039 , "cache_read_input_tokens" : 7123 , "cache_creation_input_tokens" : 7345 , "server_tool_use" : { "web_search_requests" : 1 } } }\n \ue056\n\n Web search is available on the Claude API for $10 per 1,000 searches, plus standard token costs for search-generated content. Web search results retrieved throughout a conversation are counted as input tokens, in search iterations executed during a single turn and in subsequent conversation turns.\n\n Each web search counts as one use, regardless of the number of results returned. If an error occurs during web search, the web search will not be billed.\n\n \ue09a\n Web fetch tool\n\n Web fetch usage has no additional charges beyond standard token costs:\n\n { "usage" : { "input_tokens" : 25039 , "output_tokens" : 931 , "cache_read_input_tokens" : 0 , "cache_creation_input_tokens" : 0 , "server_tool_use" : { "web_fetch_requests" : 1 } } }\n \ue056\n\n The web fetch tool is available on the Claude API at no additional cost. You only pay standard token costs for the fetched content that becomes part of your conversation context.\n\n To protect against inadvertently fetching large content that would consume excessive tokens, use the max_content_tokens parameter to set appropriate limits based on your use case and budget considerations.\n\n Example token usage for typical content:\n\n * Average web page (10 kB): ~2,500 tokens\n * Large documentation page (100 kB): ~25,000 tokens\n * Research paper PDF (500 kB): ~125,000 tokens\n\n \ue09a\n Computer use tool\n\n Computer use follows the standard tool use pricing. When using the computer use tool:\n\n System prompt overhead: The computer use beta adds 466-499 tokens to the system prompt\n\n Computer use tool token usage:\n\n Model Input tokens per tool definition\n Claude 4.x models 735 tokens\n\n\n Additional token consumption:\n\n * Screenshot images (see Vision pricing)\n * Tool execution results returned to Claude\n Session runtime replaces the Code Execution container-hour billing model when using Claude Managed Agents. You are not separately billed for container hours on top of session runtime.\n If you're also using bash or text editor tools alongside computer use, those tools have their own token costs as documented in their respective pages.\n Worked example\n\n A one-hour coding session using Claude Opus 4.8 that consumes 50,000 input tokens and 15,000 output tokens:\n\n Line item Calculation Cost\n Input tokens 50,000 \u00d7 $5 / 1,000,000 $0.25\n Output tokens 15,000 \u00d7 $25 / 1,000,000 $0.375\n Session runtime 1.0 hour \u00d7 $0.08 $0.08\n Total $0.705\n\n\n If prompt caching is active and 40,000 of the input tokens are cache reads:\n\n Line item Calculation Cost\n Uncached input tokens 10,000 \u00d7 $5 / 1,000,000 $0.05\n Cache read tokens 40,000 \u00d7 $5 \u00d7 0.1 / 1,000,000 $0.02\n Output tokens 15,000 \u00d7 $25 / 1,000,000 $0.375\n Session runtime 1.0 hour \u00d7 $0.08 $0.08\n Total $0.525\n Claude Managed Agents pricing\n\n Claude Managed Agents is billed on two dimensions: tokens and session runtime.\n\n \ue09a\n Tokens\n\n All tokens consumed by a Claude Managed Agents session are billed at the rates shown in Model pricing. Prompt caching multipliers apply identically. Web search triggered inside a session incurs the standard $10 per 1,000 searches. On Claude Platform on AWS, session token and runtime charges convert to Claude Consumption Units at the standard rate.\n\n The following Messages API modifiers do not apply to Claude Managed Agents sessions:\n\n Modifier Why it doesn't apply\n Batch API discount Sessions are stateful and interactive. There is no batch mode.\n Fast mode premium Inference speed is managed by the runtime.\n Data residency multiplier inference_geo is a Messages API request field.\n Cloud platform pricing Not available on partner-operated cloud platforms.\n\n\n \ue09a\n Session runtime\n\n SKU Rate Metering\n Session runtime $0.08 per session-hour running status duration\n\n\n Runtime is measured to the millisecond and accrues only while the session's status is running. Time spent idle (waiting for your next message or a tool confirmation), rescheduling, or terminated does not count toward runtime.\n Example calculation for processing 10,000 support tickets:\n\n * Average ~3,700 tokens per conversation\n * Using Claude Haiku 4.5 at $1/MTok input, $5/MTok output\n * Total cost: ~$37.00 per 10,000 tickets\n\n For a detailed walkthrough of this calculation, see the customer support agent guide.\n Session runtime replaces the Code Execution container-hour billing model when using Claude Managed Agents. You are not separately billed for container hours on top of session runtime.\n Additional pricing considerations\n\n \ue09a\n Cost optimization strategies\n\n When building agents with Claude:\n\n 1. Use appropriate models: Choose Haiku for simple tasks, Sonnet for most production workloads, and Opus for the most complex reasoning\n 2. Implement prompt caching: Reduce costs for repeated context\n 3. Batch operations: Use the Batch API for non-time-sensitive tasks\n 4. Monitor usage patterns: Track token consumption to identify optimization opportunities\n \ue097\n\n For high-volume agent applications, contact the enterprise sales team for custom pricing arrangements.\n\n \ue09a\n Rate limits\n\n Rate limits vary by usage tier and affect how many requests you can make:\n\n * Start tier: Entry-level limits for getting started\n * Build tier: Increased limits for growing applications\n * Scale tier: Highest standard limits for production workloads\n\n For detailed rate limit information, see Rate limits.\n\n For limits beyond the Scale tier or custom pricing arrangements, contact the sales team.\n\n \ue09a\n Volume discounts\n\n Volume discounts may be available for high-volume users. These are negotiated on a case-by-case basis.\n\n * Standard usage tiers use the pricing shown in Model pricing\n * Enterprise customers can contact sales for custom pricing\n * Academic and research discounts may be available\n\n \ue09a\n Enterprise pricing\n\n For enterprise customers with specific needs:\n\n * Custom rate limits\n * Volume discounts\n * Dedicated support\n * Custom terms\n\n Contact the sales team at sales@anthropic.com or through the Claude Console to discuss enterprise pricing options.\n\n \ue09a\n Billing and payment\n\n * Billing is based on actual monthly usage\n * All payments are in USD\n * Credit card and invoicing options available\n * Usage tracking available in the Claude Console\n\n \ue09a\n Frequently asked questions\n\n How is token usage calculated?\n\n Tokens are pieces of text that models process. As a rough estimate, 1 token is approximately 4 characters or 0.75 words in English. The exact count varies by language and content type.\n\n Are there free tiers or trials?\n\n New users receive a small amount of free credits to test the API. Contact sales for information about extended trials for enterprise evaluation.\n\n How do discounts stack?\n\n Batch API and prompt caching discounts can be combined. For example, using both features together provides significant cost savings compared to standard API calls. See prompt caching pricing for how the multipliers interact.\n\n What payment methods are accepted?\n\n Major credit cards are accepted for standard accounts. Enterprise customers can arrange invoicing and other payment methods.\n\n For additional questions about pricing, contact support@anthropic.com.\n\n Was this page helpful?\n\n \ue0fb \ue0f9\n * Model pricing\n * Cloud platform pricing\n * Claude Platform on AWS pricing\n * Inference geography\n * Private offers\n * Claude in Microsoft Foundry pricing\n * Inference geography\n * Feature-specific pricing\n * Prompt caching\n * Data residency pricing\n * Fast mode pricing\n * Batch processing\n * Long context pricing\n * Tool use pricing\n * Specific tool pricing\n * Claude Managed Agents pricing\n * Tokens\n * Session runtime\n * Worked example\n * Additional pricing considerations\n * Cost optimization strategies\n * Rate limits\n * Volume discounts\n * Enterprise pricing\n * Billing and payment\n * Frequently asked questions\n Claude Platform Docs\n\n Solutions\n\n * AI agents\n * Code modernization\n * Coding\n * Customer support\n * Education\n * Financial services\n * Government\n * Life sciences\n\n Partners\n\n * Claude on AWS\n * Claude on Google Cloud\n\n Learn\n\n * Blog\n * Courses\n * Use cases\n * Connectors\n * Customer stories\n * Engineering at Anthropic\n * Events\n * Powered by Claude\n * Service partners\n * Startups program\n\n Company\n\n * Anthropic\n * Careers\n * Economic Futures\n * Research\n * News\n * Responsible Scaling Policy\n * Security and compliance\n * Transparency\n\n Learn\n\n * Blog\n * Courses\n * Use cases\n * Connectors\n * Customer stories\n * Engineering at Anthropic\n * Events\n * Powered by Claude\n * Service partners\n * Startups program\n\n Help and security\n\n * Availability\n * Status\n * Support\n * Discord\n\n Terms and policies\n\n * Privacy policy\n * Responsible disclosure policy\n * Terms of service: Commercial\n * Terms of service: Consumer\n * Usage policy\n Worked example\n\n A one-hour coding session using Claude Opus 4.8 that consumes 50,000 input tokens and 15,000 output tokens:\n\n Line item Calculation Cost\n Input tokens 50,000 \u00d7 $5 / 1,000,000 $0.25\n Output tokens 15,000 \u00d7 $25 / 1,000,000 $0.375\n Session runtime 1.0 hour \u00d7 $0.08 $0.08\n Total $0.705\n\n\n If prompt caching is active and 40,000 of the input tokens are cache reads:\n\n Line item Calculation Cost\n Uncached input tokens 10,000 \u00d7 $5 / 1,000,000 $0.05\n Cache read tokens 40,000 \u00d7 $5 \u00d7 0.1 / 1,000,000 $0.02\n Output tokens 15,000 \u00d7 $25 / 1,000,000 $0.375\n Session runtime 1.0 hour \u00d7 $0.08 $0.08\n Total $0.525\n\n \ue08f\n\n Example calculation for processing 10,000 support tickets:\n\n * Average ~3,700 tokens per conversation\n * Using Claude Haiku 4.5 at $1/MTok input, $5/MTok output\n * Total cost: ~$37.00 per 10,000 tickets\n\n For a detailed walkthrough of this calculation, see the customer support agent guide.\n\n \ue09a\n Additional pricing considerations\n\n \ue09a\n Cost optimization strategies\n\n When building agents with Claude:\n\n 1. Use appropriate models: Choose Haiku for simple tasks, Sonnet for most production workloads, and Opus for the most complex reasoning\n 2. Implement prompt caching: Reduce costs for repeated context\n 3. Batch operations: Use the Batch API for non-time-sensitive tasks\n 4. Monitor usage patterns: Track token consumption to identify optimization opportunities\n \ue097\n\n For high-volume agent applications, contact the enterprise sales team for custom pricing arrangements.\n\n \ue09a\n Rate limits\n\n Rate limits vary by usage tier and affect how many requests you can make:\n\n * Start tier: Entry-level limits for getting started\n * Build tier: Increased limits for growing applications\n * Scale tier: Highest standard limits for production workloads\n\n For detailed rate limit information, see Rate limits.\n\n For limits beyond the Scale tier or custom pricing arrangements, contact the sales team.\n\n \ue09a\n Volume discounts\n\n Volume discounts may be available for high-volume users. These are negotiated on a case-by-case basis.\n\n * Standard usage tiers use the pricing shown in Model pricing\n * Enterprise customers can contact sales for custom pricing\n * Academic and research discounts may be available\n\n \ue09a\n Enterprise pricing\n\n For enterprise customers with specific needs:\n\n * Custom rate limits\n * Volume discounts\n * Dedicated support\n * Custom terms\n\n Contact the sales team at sales@anthropic.com or through the Claude Console to discuss enterprise pricing options.\n\n \ue09a\n Billing and payment\n\n * Billing is based on actual monthly usage\n * All payments are in USD\n * Credit card and invoicing options available\n * Usage tracking available in the Claude Console\n\n \ue09a\n Frequently asked questions\n\n How is token usage calculated?\n\n Tokens are pieces of text that models process. As a rough estimate, 1 token is approximately 4 characters or 0.75 words in English. The exact count varies by language and content type.\n\n Are there free tiers or trials?\n\n New users receive a small amount of free credits to test the API. Contact sales for information about extended trials for enterprise evaluation.\n\n How do discounts stack?\n\n Batch API and prompt caching discounts can be combined. For example, using both features together provides significant cost savings compared to standard API calls. See prompt caching pricing for how the multipliers interact.\n\n What payment methods are accepted?\n\n Major credit cards are accepted for standard accounts. Enterprise customers can arrange invoicing and other payment methods.\n\n For additional questions about pricing, contact support@anthropic.com.\n\n Was this page helpful?\n\n \ue0fb \ue0f9\n * Model pricing\n * Cloud platform pricing\n * Claude Platform on AWS pricing\n * Inference geography\n * Private offers\n * Claude in Microsoft Foundry pricing\n * Inference geography\n * Feature-specific pricing\n * Prompt caching\n * Data residency pricing\n * Fast mode pricing\n * Batch processing\n * Long context pricing\n * Tool use pricing\n * Specific tool pricing\n * Claude Managed Agents pricing\n * Tokens\n * Session runtime\n * Worked example\n * Additional pricing considerations\n * Cost optimization strategies\n * Rate limits\n * Volume discounts\n * Enterprise pricing\n * Billing and payment\n * Frequently asked questions\n Claude Platform Docs\n\n Solutions\n\n * AI agents\n * Code modernization\n * Coding\n * Customer support\n * Education\n * Financial services\n * Government\n * Life sciences\n\n Partners\n\n * Claude on AWS\n * Claude on Google Cloud\n\n Learn\n\n * Blog\n * Courses\n * Use cases\n * Connectors\n * Customer stories\n * Engineering at Anthropic\n * Events\n * Powered by Claude\n * Service partners\n * Startups program\n\n Company\n\n * Anthropic\n * Careers\n * Economic Futures\n * Research\n * News\n * Responsible Scaling Policy\n * Security and compliance\n * Transparency\n\n Learn\n\n * Blog\n * Courses\n * Use cases\n * Connectors\n * Customer stories\n * Engineering at Anthropic\n * Events\n * Powered by Claude\n * Service partners\n * Startups program\n\n Help and security\n\n * Availability\n * Status\n * Support\n * Discord\n\n Terms and policies\n\n * Privacy policy\n * Responsible disclosure policy\n * Terms of service: Commercial\n * Terms of service: Consumer\n * Usage policy", "preview_url": "https://changedetection.benfou.fr/preview/f10610bf-32e8-4762-a1c2-e9481624471d", "screenshot": null, "timestamp_from": null, "timestamp_to": null, "triggered_text": "", "llm_summary": null, "llm_intent": null, "uuid": "f10610bf-32e8-4762-a1c2-e9481624471d", "watch_mime_type": "text/html; charset=utf-8", "watch_tag": "IA", "watch_title": "API Claude", "watch_url": "https://platform.claude.com/docs/en/about-claude/pricing", "watch_uuid": "f10610bf-32e8-4762-a1c2-e9481624471d", "notification_urls": ["jsons://n8n.benfou.fr/webhook/c8bb38ef-b9d8-45d0-837d-c239411db01b/?title={{ watch_title|urlencode }}"], "notification_title": "Changement d\u00e9tect\u00e9 - {{watch_title}}", "notification_body": "{{watch_url}} a chang\u00e9e.\r\n---\r\n{{diff}}\r\n---\r\n", "notification_format": "htmlcolor", "_llm_result": null, "_llm_intent": "", "_llm_change_summary": "", "edit_url": "https://changedetection.benfou.fr/edit/f10610bf-32e8-4762-a1c2-e9481624471d"}}]
Thu Jul 16 07:12:42 2026 - SENDING - [{"title": "Changement d\u00e9tect\u00e9 - API OpenIA", "body": "https://developers.openai.com/api/docs/changelog a chang\u00e9e. \n--- \n    Get started \n \n      * Overview \n      * Quickstart \n      * Models \n      * Pricing \n      * SDKs and CLI \n          + OpenAI SDK \n          + Agents SDK \n          + OpenAI CLI \n      * Latest: GPT-5.5 \n      * Prompt guidance \n \n    Core concepts \n \n      * Text generation \n      * Code generation \n      * Images and vision \n      * Audio and speech \n      * Structured output \n      * Function calling \n      * Responses API \n      * Using tools \n \n    Agents SDK \n \n      * Overview \n      * Quickstart \n      * Agent definitions \n      * Models and providers \n      * Running agents \n      * Sandbox agents \n      * Orchestration \n      * Guardrails \n      * Results and state \n      * Integrations and observability \n      * Evaluate agent workflows \n      * Voice agents \n      * ChatKit \n          + Overview \n          + Customize \n          + Widgets \n          + Actions \n          + Advanced integrations \n \n    Tools \n \n      * Web search \n      * MCP and Connectors \n          + Secure MCP Tunnel \n      * Skills \n      * Shell \n      * Computer use \n      * File search and retrieval \n          + File search \n          + Retrieval \n      * Tool search \n      * More tools \n          + Apply Patch \n          + Local shell \n          + Image generation \n          + Code interpreter \n \n    Run and scale \n \n      * Conversation state \n      * Background mode \n      * Streaming \n      * WebSocket mode \n      * Webhooks \n      * File inputs \n      * Context management \n          + Compaction \n          + Counting tokens \n          + Prompt caching \n      * Prompting \n          + Overview \n          + Prompt engineering \n          + Citation formatting \n          + Migration guide \n      * Reasoning \n          + Reasoning models \n          + Reasoning best practices \n \n    Evaluation \n \n      * Red teaming \n \n    Realtime and audio \n \n      * Overview \n      * Voice agents \n      * Live translation \n      * Transcription \n          + Realtime transcription \n          + Speech to text \n      * Speech generation \n      * Realtime prompting guide \n      * Connection methods \n          + WebRTC \n          + WebSocket \n          + SIP \n      * Realtime sessions \n          + Managing conversations \n          + Voice activity detection \n          + Realtime with tools \n          + Webhooks and server-side controls \n          + Managing costs \n \n    Specialized models \n \n      * Image generation \n      * Video generation \n      * Deep research \n      * Embeddings \n      * Moderation \n \n    Going live \n \n      * Production best practices \n      * Workload identity federation \n          + Overview \n          + Kubernetes \n          + AWS \n          + Microsoft Azure \n          + Google Cloud \n          + GitHub Actions \n          + SPIFFE \n      * Deployment checklist \n      * Amazon Bedrock \n      * Latency optimization \n          + Overview \n          + Predicted Outputs \n          + Priority processing \n      * Cost optimization \n          + Overview \n          + Batch \n          + Flex processing \n      * Accuracy optimization \n      * Safety \n          + Safety best practices \n          + Safety checks \n          + Cybersecurity checks \n          + Under 18 API Guidance \n \n    Legacy APIs \n \n      * Agent Builder \n          + Overview \n          + Migration guide \n          + Node reference \n          + Safety in building agents \n      * Evals \n          + Getting started \n          + Working with evals \n          + Prompt optimizer \n          + External models \n          + Best practices \n          + Graders \n      * Fine-tuning \n          + Optimization cycle \n          + Supervised fine-tuning \n          + Vision fine-tuning \n          + Direct preference optimization \n          + Reinforcement fine-tuning \n          + RFT use cases \n          + Best practices \n      * Assistants API \n          + Migration guide \n          + Deep dive \n          + Tools \n \n    Resources \n \n      * Terms and policies \n      * Changelog \n      * Your data \n      * Permissions \n      * Rate limits \n      * IP egress ranges \n      * Admin APIs \n      * Deprecations \n      * MCP for deep research \n      * Developer mode \n      * ChatGPT Actions \n          + Introduction \n          + Getting started \n          + Actions library \n          + Authentication \n          + Production \n          + Data retrieval \n          + Sending files \n    Primary navigation \n      API Codex ChatGPT Docs Use cases Resources Resources \n            Search docs \n \n              Suggested \n \n                response_formatreasoning_effortstreamingtools \n        Overview Models Agents Tools Voice & Audio Production API reference \n            Overview Models Agents Tools Voice & Audio Production API reference Docs section Overview \n            * Home \n \n          Get started \n \n            * Quickstart \n            * Using GPT-5.6 \n            * Key concepts \n \n          Core concepts \n \n            * Responses API \n            * Conversation state \n            * Background mode \n            * Streaming \n            * WebSocket mode \n            * Multi-agent \n            * Webhooks \n            * File inputs \n            * Compaction \n            * Counting tokens \n \n          SDKs and CLI \n \n            * OpenAI SDK \n            * OpenAI CLI \n \n          Resources \n \n            * Changelog \n            * Deprecations \n            * Supported countries \n            * OpenAI Crawlers \n            * Terms and policies \n \n          Legacy APIs \n \n            * Agent Builder \n                + Overview \n                + Migration guide \n                + Node reference \n                + Safety in building agents \n            * Evals \n                + Getting started \n                + Working with evals \n                + Prompt optimizer \n                + External models \n                + Best practices \n                + Graders \n            * Fine-tuning \n                + Optimization cycle \n                + Supervised fine-tuning \n                + Vision fine-tuning \n                + Direct preference optimization \n                + Reinforcement fine-tuning \n                + RFT use cases \n                + Best practices \n            * Assistants API \n                + Migration guide \n                + Deep dive \n                + Tools \n            * Model catalog \n \n          Choose a model \n \n            * Pricing \n            * Model selection \n \n          Text and code \n \n            * Text generation \n            * Code generation \n            * Structured output \n \n          Prompting \n \n            * Overview \n            * Prompt engineering \n            * Citation formatting \n            * Migration guide \n            * Prompt generation \n            * Frontend prompting \n \n          Reasoning \n \n            * Reasoning models \n            * Reasoning best practices \n \n          Images and video \n \n            * Images and vision \n            * Image generation \n            * Video generation \n \n          Realtime and audio \n \n            * Audio and speech \n            * Overview \n            * Voice agents \n \n          Specialized models \n \n            * Deep research \n            * Embeddings \n            * Moderation \n            * Overview \n \n          Agents SDK \n \n            * Quickstart \n            * Agent definitions \n            * Models and providers \n            * Running agents \n            * Sandbox agents \n            * Orchestration \n            * Guardrails \n            * Results and state \n            * Integrations and observability \n            * Evaluate agent workflows \n \n          ChatKit \n \n            * Overview \n            * Customize \n            * Widgets \n            * Actions \n            * Advanced integrations \n            * Overview \n            * Function calling \n \n          Search and retrieval \n \n            * Web search \n            * File search \n            * Retrieval \n \n          Connect tools and data \n \n            * MCP and Connectors \n            * Secure MCP Tunnel \n \n          Build tool workflows \n \n            * Skills \n            * Tool search \n            * Programmatic tool calling \n \n          Computer and code \n \n            * Shell \n            * Computer use \n            * Apply Patch \n            * Local shell \n            * Code interpreter \n \n          Media \n \n            * Image generation \n            * Overview \n \n          Get started \n \n            * Voice agents \n            * Live translation \n            * Realtime prompting guide \n \n          Audio \n \n            * Audio and speech \n            * Realtime transcription \n            * Speech to text \n            * Speech generation \n \n          Connection methods \n \n            * WebRTC \n            * WebSocket \n            * SIP \n \n          Sessions and operations \n \n            * Managing conversations \n            * Voice activity detection \n            * Realtime with tools \n            * Webhooks and server-side controls \n            * Managing costs \n \n          Go live \n \n            * Production best practices \n            * Deployment checklist \n \n          Performance and quality \n \n            * Latency optimization \n            * Predicted Outputs \n            * Priority processing \n            * Accuracy optimization \n \n          Cost and throughput \n \n            * Cost optimization \n            * Prompt caching \n            * Batch \n            * Flex processing \n \n          Safety and governance \n \n            * Safety best practices \n            * Red teaming \n            * Safety checks \n                + Cybersecurity checks \n                + Under 18 API Guidance \n            * Your data \n            * Permissions \n \n          Infrastructure and access \n \n            * Private Link \n            * Workload identity federation \n                + Kubernetes \n                + AWS \n                + Microsoft Azure \n                + Google Cloud \n                + GitHub Actions \n                + SPIFFE \n            * IP egress ranges \n            * Amazon Bedrock \n \n          Operations \n \n            * Rate limits \n            * Admin APIs \n            * Error codes \n        Docs Use cases \n            Docs Use cases Docs section Docs \n        Apps SDK Workspace Agents Commerce Ads \n            Apps SDK Workspace Agents Commerce Ads Docs section Select... \n            * Home \n            * Quickstart \n \n          Core Concepts \n \n            * MCP Apps in ChatGPT \n            * MCP Server \n            * UX principles \n            * UI guidelines \n \n          Plan \n \n            * Research use cases \n            * Define tools \n            * Design components \n \n          Build \n \n            * Set up your server \n            * Build your ChatGPT UI \n            * Authenticate users \n            * Manage state \n            * Monetize your app \n            * Examples \n \n          Deploy \n \n            * Deploy your app \n            * Connect from ChatGPT \n            * Test your integration \n            * Prepare your app \n \n          Conversion apps \n \n            * Restaurant reservation spec \n            * Product checkout spec \n \n          Guides \n \n            * Optimize Metadata \n            * Security & Privacy \n            * Troubleshooting \n \n          Resources \n \n            * Changelog \n            * App guidelines \n            * Reference \n            * Home \n \n          Get started \n \n            * Trigger workspace agent runs \n            * Authenticate with Workspace Agent access tokens \n            * Home \n \n          Guides \n \n            * Get started \n            * Best practices \n \n          File Upload \n \n            * Overview \n            * Products \n \n          API \n \n            * Overview \n            * Feeds \n            * Products \n            * Promotions \n            * Ads Overview \n \n          Measurement \n \n            * JavaScript Pixel \n            * Image tag \n            * Conversions API \n            * Supported events \n \n          Advertiser API \n \n            * Overview \n            * Quickstart \n            * Product feeds \n            * Campaign Targeting \n \n          API Reference \n \n            * Authentication \n            * Campaigns \n            * Ad Groups \n            * Ads \n            * Ad Account \n            * Insights \n            * Files \n        Overview Features Configuration Developers Security Administration Use Cases Resources \n            Overview Features Configuration Developers Security Administration Use Cases Resources Docs section Overview \n            * Home \n \n          Get started \n \n            * Quickstart \n            * Use ChatGPT \n            * Get started with Work \n            * Import from another agent \n \n          Foundations \n \n            * Prompting \n            * Personalize ChatGPT \n            * Skills & Plugins \n            * Permissions \n \n          Explore \n \n            * What's new \n            * Models \n            * Pricing \n            * Glossary \n \n          Available on \n \n            * ChatGPT desktop app \n            * ChatGPT on the web \n            * Codex CLI \n            * Codex IDE extension \n            * Codex cloud \n \n          Releases \n \n            * Changelog \n            * Feature Maturity \n            * Open Source \n            * Overview \n \n          Workflows \n \n            * Projects, chats, and tasks \n            * Sites \n            * Visualizations \n            * Scheduled tasks \n            * Long-running work \n            * Notifications \n            * Pets \n \n          Capabilities \n \n            * Browser \n            * Computer use \n            * Plugins \n            * Web search \n            * Image generation \n            * Image inputs \n            * Appshots \n            * Chrome extension \n            * Work with files \n \n          Reference \n \n            * Commands \n            * Slash commands \n            * Settings \n            * Troubleshooting \n            * Overview \n \n          Customization \n \n            * Overview \n            * Memories \n            * Chronicle \n \n          Config file \n \n            * Config Basics \n            * Advanced Config \n            * Config Reference \n            * Environment Variables \n            * Sample Config \n \n          Agent configuration \n \n            * AGENTS.md \n            * Subagents \n            * Speed \n            * Rules \n \n          Extend ChatGPT and Codex \n \n            * Record & Replay \n            * MCP \n \n          Windows \n \n            * Desktop app \n            * Windows sandbox \n            * WSL \n            * Overview \n \n          Development workflows \n \n            * Code review \n            * Integrated terminal \n \n          Extend and automate \n \n            * Build skills \n            * Build plugins \n            * Build an app \n            * Submit plugins \n            * Hooks \n \n          Environments \n \n            * Modes \n            * Local environments \n            * Cloud environment \n            * Git worktrees \n \n          Build with Codex \n \n            * Codex SDK \n            * App Server \n            * MCP Server \n            * GitHub Action \n            * Non-interactive mode \n \n          Third-party integrations \n \n            * GitHub \n            * Slack \n            * Linear \n \n          Reference \n \n            * CLI customization \n            * Developer commands \n            * Developer settings \n            * Overview \n \n          Permissions \n \n            * Profiles \n            * Sandboxing \n            * Auto-review \n            * Agent approvals & security \n            * Internet access \n \n          Codex Security \n \n            * Overview \n            * FAQ \n            * Codex Security plugin \n                + Quickstart \n                + Run a security scan \n                + Run a deep scan \n                + Review code changes \n                + Triage a backlog \n                + Fix findings \n                + Export and track findings \n                + Changelog \n            * Codex Security cloud setup \n            * Improving the threat model \n \n          Safety \n \n            * Cyber Safety \n            * Overview \n \n          Getting started \n \n            * Admin rollout guide \n            * ChatGPT Work Admin FAQ \n \n          Identity and authentication \n \n            * Authentication overview \n            * Access tokens \n \n          Workspace access, policy, and models \n \n            * Groups and provisioning \n            * Roles and workspace permissions \n            * Managed configuration \n            * HIPAA configuration \n            * Workspace model availability \n \n          Plugin and connector controls \n \n            * Plugin controls \n            * Skill controls \n \n          Usage, governance, and compliance \n \n            * Governance \n            * Workspace analytics \n            * Analytics API \n            * Compliance API and audit events \n \n          Deployment and model providers \n \n            * Windows app deployment \n            * Remote connections \n            * Amazon Bedrock \n            * Explore use cases \n            * Collections \n            * Home \n            * Videos \n            * Showcase \n            * OpenAI Academy \n            * Online trainings \n \n          Community \n \n            * Codex Ambassadors \n            * Codex for Students \n            * Codex for Open Source \n            * Meetups \n \n          Blog \n \n            * Company blog \n            * Developer blog \n            * Explore use cases \n            * Collections \n            * Home \n            * Videos \n            * Showcase \n            * OpenAI Academy \n            * Online trainings \n \n          Community \n \n            * Codex Ambassadors \n            * Codex for Students \n            * Codex for Open Source \n            * Meetups \n \n          Blog \n \n            * Company blog \n            * Developer blog \n        Showcase Blog Cookbook Learn Community \n            Showcase Blog Cookbook Learn Community Docs section Select... \n            * All posts \n \n          Recent \n \n            * Mastering remote engineering work from your phone \n            * Making private MCP servers reachable without making them public \n            * How Perplexity Brought Voice Search to Millions Using the Realtime API \n            * Designing delightful frontends with GPT-5.4 \n            * From prompts to products: One year of Responses \n \n          Topics \n \n            * General \n            * API \n            * Apps SDK \n            * Audio \n            * Codex \n            * Home \n \n          Topics \n \n            * Agents \n            * Evals \n            * Multimodal \n            * Text \n            * Guardrails \n            * Optimization \n            * ChatGPT \n            * Codex \n            * gpt-oss \n \n          Contribute \n \n            * Cookbook on GitHub \n            * Home \n            * OpenAI Developers plugin \n            * Docs MCP \n \n          Categories \n \n            * Demo apps \n            * Videos \n \n          Topics \n \n            * Agents \n            * Audio & Voice \n            * Computer Use \n            * Codex \n            * Evals \n            * gpt-oss \n            * Fine-tuning \n            * Image generation \n            * Scaling \n            * Tools \n            * Video generation \n            * Community \n \n          Programs \n \n            * Codex Ambassadors \n            * Codex for Students \n            * Codex for Open Source \n            * OpenAI for Startups \n \n          Events \n \n            * Meetups \n \n          Spaces \n \n            * Developer Forum \n            * Discord \n            * Reddit \n            * X \n          API Dashboard \n          Try ChatGPT \n            * Codex Micro \n---", "url": "jsons://n8n.benfou.fr/webhook/c8bb38ef-b9d8-45d0-837d-c239411db01b/?title=API%20OpenIA&avatar_url=https://raw.githubusercontent.com/dgtlmoon/changedetection.io/master/changedetectionio/static/images/avatar-256x256.png&format=html", "original_context": {"base_url": "https://changedetection.benfou.fr/", "change_datetime": "2026-07-16 07:12:16 CEST", "current_snapshot": " * Codex Micro", "diff": "@changed_PLACEMARKER_OPEN Get started@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Overview@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Quickstart@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Models@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Pricing@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * SDKs and CLI@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN + OpenAI SDK@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN + Agents SDK@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN + OpenAI CLI@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Latest: GPT-5.5@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Prompt guidance@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Core concepts@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Text generation@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Code generation@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Images and vision@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Audio and speech@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Structured output@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Function calling@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Responses API@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Using tools@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Agents SDK@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Overview@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Quickstart@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Agent definitions@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Models and providers@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Running agents@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Sandbox agents@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Orchestration@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Guardrails@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Results and state@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Integrations and observability@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Evaluate agent workflows@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Voice agents@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * ChatKit@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN + Overview@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN + Customize@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN + Widgets@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN + Actions@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN + Advanced integrations@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Tools@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Web search@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * MCP and Connectors@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN + Secure MCP Tunnel@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Skills@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Shell@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Computer use@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * File search and retrieval@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN + File search@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN + Retrieval@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Tool search@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * More tools@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN + Apply Patch@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN + Local shell@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN + Image generation@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN + Code interpreter@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Run and scale@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Conversation state@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Background mode@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Streaming@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * WebSocket mode@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Webhooks@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * File inputs@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Context management@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN + Compaction@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN + Counting tokens@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN + Prompt caching@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Prompting@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN + Overview@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN + Prompt engineering@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN + Citation formatting@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN + Migration guide@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Reasoning@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN + Reasoning models@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN + Reasoning best practices@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Evaluation@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Red teaming@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Realtime and audio@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Overview@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Voice agents@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Live translation@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Transcription@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN + Realtime transcription@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN + Speech to text@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Speech generation@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Realtime prompting guide@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Connection methods@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN + WebRTC@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN + WebSocket@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN + SIP@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Realtime sessions@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN + Managing conversations@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN + Voice activity detection@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN + Realtime with tools@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN + Webhooks and server-side controls@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN + Managing costs@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Specialized models@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Image generation@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Video generation@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Deep research@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Embeddings@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Moderation@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Going live@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Production best practices@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Workload identity federation@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN + Overview@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN + Kubernetes@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN + AWS@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN + Microsoft Azure@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN + Google Cloud@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN + GitHub Actions@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN + SPIFFE@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Deployment checklist@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Amazon Bedrock@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Latency optimization@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN + Overview@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN + Predicted Outputs@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN + Priority processing@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Cost optimization@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN + Overview@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN + Batch@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN + Flex processing@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Accuracy optimization@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Safety@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN + Safety best practices@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN + Safety checks@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN + Cybersecurity checks@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN + Under 18 API Guidance@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Legacy APIs@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Agent Builder@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN + Overview@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN + Migration guide@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN + Node reference@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN + Safety in building agents@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Evals@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN + Getting started@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN + Working with evals@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN + Prompt optimizer@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN + External models@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN + Best practices@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN + Graders@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Fine-tuning@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN + Optimization cycle@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN + Supervised fine-tuning@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN + Vision fine-tuning@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN + Direct preference optimization@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN + Reinforcement fine-tuning@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN + RFT use cases@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN + Best practices@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Assistants API@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN + Migration guide@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN + Deep dive@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN + Tools@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Resources@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Terms and policies@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Changelog@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Your data@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Permissions@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Rate limits@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * IP egress ranges@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Admin APIs@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Deprecations@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * MCP for deep research@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Developer mode@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * ChatGPT Actions@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN + Introduction@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN + Getting started@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN + Actions library@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN + Authentication@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN + Production@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN + Data retrieval@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN + Sending files@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Primary navigation@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN API Codex ChatGPT Docs Use cases Resources Resources@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Search docs@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Suggested@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN response_formatreasoning_effortstreamingtools@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Overview Models Agents Tools Voice & Audio Production API reference@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Overview Models Agents Tools Voice & Audio Production API reference Docs section Overview@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Home@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Get started@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Quickstart@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Using GPT-5.6@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Key concepts@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Core concepts@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Responses API@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Conversation state@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Background mode@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Streaming@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * WebSocket mode@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Multi-agent@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Webhooks@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * File inputs@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Compaction@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Counting tokens@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN SDKs and CLI@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * OpenAI SDK@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * OpenAI CLI@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Resources@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Changelog@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Deprecations@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Supported countries@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * OpenAI Crawlers@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Terms and policies@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Legacy APIs@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Agent Builder@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN + Overview@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN + Migration guide@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN + Node reference@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN + Safety in building agents@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Evals@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN + Getting started@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN + Working with evals@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN + Prompt optimizer@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN + External models@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN + Best practices@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN + Graders@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Fine-tuning@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN + Optimization cycle@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN + Supervised fine-tuning@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN + Vision fine-tuning@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN + Direct preference optimization@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN + Reinforcement fine-tuning@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN + RFT use cases@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN + Best practices@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Assistants API@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN + Migration guide@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN + Deep dive@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN + Tools@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Model catalog@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Choose a model@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Pricing@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Model selection@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Text and code@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Text generation@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Code generation@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Structured output@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Prompting@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Overview@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Prompt engineering@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Citation formatting@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Migration guide@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Prompt generation@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Frontend prompting@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Reasoning@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Reasoning models@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Reasoning best practices@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Images and video@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Images and vision@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Image generation@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Video generation@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Realtime and audio@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Audio and speech@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Overview@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Voice agents@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Specialized models@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Deep research@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Embeddings@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Moderation@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Overview@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Agents SDK@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Quickstart@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Agent definitions@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Models and providers@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Running agents@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Sandbox agents@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Orchestration@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Guardrails@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Results and state@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Integrations and observability@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Evaluate agent workflows@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN ChatKit@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Overview@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Customize@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Widgets@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Actions@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Advanced integrations@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Overview@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Function calling@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Search and retrieval@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Web search@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * File search@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Retrieval@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Connect tools and data@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * MCP and Connectors@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Secure MCP Tunnel@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Build tool workflows@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Skills@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Tool search@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Programmatic tool calling@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Computer and code@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Shell@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Computer use@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Apply Patch@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Local shell@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Code interpreter@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Media@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Image generation@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Overview@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Get started@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Voice agents@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Live translation@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Realtime prompting guide@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Audio@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Audio and speech@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Realtime transcription@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Speech to text@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Speech generation@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Connection methods@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * WebRTC@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * WebSocket@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * SIP@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Sessions and operations@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Managing conversations@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Voice activity detection@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Realtime with tools@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Webhooks and server-side controls@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Managing costs@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Go live@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Production best practices@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Deployment checklist@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Performance and quality@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Latency optimization@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Predicted Outputs@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Priority processing@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Accuracy optimization@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Cost and throughput@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Cost optimization@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Prompt caching@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Batch@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Flex processing@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Safety and governance@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Safety best practices@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Red teaming@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Safety checks@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN + Cybersecurity checks@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN + Under 18 API Guidance@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Your data@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Permissions@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Infrastructure and access@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Private Link@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Workload identity federation@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN + Kubernetes@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN + AWS@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN + Microsoft Azure@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN + Google Cloud@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN + GitHub Actions@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN + SPIFFE@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * IP egress ranges@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Amazon Bedrock@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Operations@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Rate limits@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Admin APIs@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Error codes@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Docs Use cases@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Docs Use cases Docs section Docs@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Apps SDK Workspace Agents Commerce Ads@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Apps SDK Workspace Agents Commerce Ads Docs section Select...@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Home@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Quickstart@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Core Concepts@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * MCP Apps in ChatGPT@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * MCP Server@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * UX principles@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * UI guidelines@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Plan@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Research use cases@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Define tools@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Design components@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Build@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Set up your server@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Build your ChatGPT UI@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Authenticate users@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Manage state@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Monetize your app@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Examples@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Deploy@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Deploy your app@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Connect from ChatGPT@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Test your integration@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Prepare your app@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Conversion apps@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Restaurant reservation spec@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Product checkout spec@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Guides@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Optimize Metadata@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Security & Privacy@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Troubleshooting@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Resources@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Changelog@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * App guidelines@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Reference@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Home@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Get started@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Trigger workspace agent runs@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Authenticate with Workspace Agent access tokens@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Home@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Guides@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Get started@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Best practices@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN File Upload@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Overview@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Products@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN API@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Overview@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Feeds@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Products@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Promotions@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Ads Overview@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Measurement@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * JavaScript Pixel@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Image tag@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Conversions API@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Supported events@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Advertiser API@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Overview@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Quickstart@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Product feeds@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Campaign Targeting@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN API Reference@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Authentication@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Campaigns@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Ad Groups@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Ads@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Ad Account@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Insights@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Files@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Overview Features Configuration Developers Security Administration Use Cases Resources@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Overview Features Configuration Developers Security Administration Use Cases Resources Docs section Overview@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Home@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Get started@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Quickstart@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Use ChatGPT@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Get started with Work@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Import from another agent@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Foundations@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Prompting@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Personalize ChatGPT@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Skills & Plugins@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Permissions@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Explore@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * What's new@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Models@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Pricing@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Glossary@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Available on@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * ChatGPT desktop app@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * ChatGPT on the web@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Codex CLI@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Codex IDE extension@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Codex cloud@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Releases@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Changelog@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Feature Maturity@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Open Source@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Overview@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Workflows@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Projects, chats, and tasks@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Sites@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Visualizations@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Scheduled tasks@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Long-running work@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Notifications@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Pets@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Capabilities@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Browser@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Computer use@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Plugins@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Web search@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Image generation@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Image inputs@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Appshots@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Chrome extension@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Work with files@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Reference@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Commands@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Slash commands@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Settings@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Troubleshooting@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Overview@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Customization@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Overview@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Memories@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Chronicle@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Config file@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Config Basics@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Advanced Config@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Config Reference@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Environment Variables@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Sample Config@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Agent configuration@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * AGENTS.md@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Subagents@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Speed@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Rules@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Extend ChatGPT and Codex@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Record & Replay@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * MCP@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Windows@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Desktop app@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Windows sandbox@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * WSL@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Overview@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Development workflows@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Code review@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Integrated terminal@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Extend and automate@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Build skills@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Build plugins@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Build an app@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Submit plugins@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Hooks@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Environments@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Modes@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Local environments@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Cloud environment@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Git worktrees@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Build with Codex@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Codex SDK@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * App Server@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * MCP Server@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * GitHub Action@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Non-interactive mode@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Third-party integrations@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * GitHub@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Slack@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Linear@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Reference@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * CLI customization@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Developer commands@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Developer settings@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Overview@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Permissions@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Profiles@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Sandboxing@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Auto-review@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Agent approvals & security@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Internet access@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Codex Security@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Overview@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * FAQ@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Codex Security plugin@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN + Quickstart@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN + Run a security scan@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN + Run a deep scan@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN + Review code changes@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN + Triage a backlog@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN + Fix findings@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN + Export and track findings@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN + Changelog@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Codex Security cloud setup@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Improving the threat model@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Safety@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Cyber Safety@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Overview@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Getting started@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Admin rollout guide@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * ChatGPT Work Admin FAQ@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Identity and authentication@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Authentication overview@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Access tokens@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Workspace access, policy, and models@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Groups and provisioning@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Roles and workspace permissions@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Managed configuration@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * HIPAA configuration@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Workspace model availability@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Plugin and connector controls@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Plugin controls@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Skill controls@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Usage, governance, and compliance@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Governance@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Workspace analytics@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Analytics API@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Compliance API and audit events@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Deployment and model providers@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Windows app deployment@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Remote connections@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Amazon Bedrock@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Explore use cases@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Collections@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Home@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Videos@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Showcase@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * OpenAI Academy@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Online trainings@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Community@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Codex Ambassadors@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Codex for Students@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Codex for Open Source@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Meetups@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Blog@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Company blog@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Developer blog@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Explore use cases@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Collections@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Home@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Videos@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Showcase@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * OpenAI Academy@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Online trainings@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Community@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Codex Ambassadors@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Codex for Students@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Codex for Open Source@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Meetups@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Blog@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Company blog@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Developer blog@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Showcase Blog Cookbook Learn Community@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Showcase Blog Cookbook Learn Community Docs section Select...@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * All posts@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Recent@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Mastering remote engineering work from your phone@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Making private MCP servers reachable without making them public@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * How Perplexity Brought Voice Search to Millions Using the Realtime API@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Designing delightful frontends with GPT-5.4@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * From prompts to products: One year of Responses@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Topics@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * General@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * API@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Apps SDK@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Audio@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Codex@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Home@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Topics@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Agents@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Evals@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Multimodal@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Text@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Guardrails@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Optimization@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * ChatGPT@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Codex@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * gpt-oss@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Contribute@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Cookbook on GitHub@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Home@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * OpenAI Developers plugin@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Docs MCP@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Categories@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Demo apps@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Videos@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Topics@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Agents@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Audio & Voice@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Computer Use@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Codex@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Evals@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * gpt-oss@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Fine-tuning@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Image generation@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Scaling@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Tools@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Video generation@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Community@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Programs@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Codex Ambassadors@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Codex for Students@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Codex for Open Source@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * OpenAI for Startups@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Events@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Meetups@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Spaces@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Developer Forum@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Discord@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Reddit@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * X@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN API Dashboard@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Try ChatGPT@changed_PLACEMARKER_CLOSED\n@changed_into_PLACEMARKER_OPEN * Codex Micro@changed_into_PLACEMARKER_CLOSED", "diff_clean": "", "diff_added": "", "diff_added_clean": "", "diff_full": "", "diff_full_clean": "", "diff_patch": "", "diff_removed": "", "diff_removed_clean": "", "diff_changed_from": "", "diff_changed_to": "", "diff_url": "https://changedetection.benfou.fr/diff/40d1d040-eab5-44c8-a830-96ce77c471dd", "raw_diff": "@changed_PLACEMARKER_OPEN Get started@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Overview@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Quickstart@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Models@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Pricing@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * SDKs and CLI@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN + OpenAI SDK@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN + Agents SDK@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN + OpenAI CLI@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Latest: GPT-5.5@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Prompt guidance@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Core concepts@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Text generation@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Code generation@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Images and vision@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Audio and speech@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Structured output@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Function calling@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Responses API@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Using tools@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Agents SDK@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Overview@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Quickstart@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Agent definitions@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Models and providers@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Running agents@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Sandbox agents@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Orchestration@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Guardrails@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Results and state@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Integrations and observability@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Evaluate agent workflows@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Voice agents@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * ChatKit@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN + Overview@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN + Customize@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN + Widgets@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN + Actions@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN + Advanced integrations@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Tools@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Web search@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * MCP and Connectors@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN + Secure MCP Tunnel@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Skills@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Shell@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Computer use@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * File search and retrieval@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN + File search@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN + Retrieval@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Tool search@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * More tools@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN + Apply Patch@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN + Local shell@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN + Image generation@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN + Code interpreter@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Run and scale@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Conversation state@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Background mode@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Streaming@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * WebSocket mode@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Webhooks@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * File inputs@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Context management@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN + Compaction@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN + Counting tokens@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN + Prompt caching@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Prompting@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN + Overview@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN + Prompt engineering@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN + Citation formatting@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN + Migration guide@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Reasoning@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN + Reasoning models@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN + Reasoning best practices@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Evaluation@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Red teaming@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Realtime and audio@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Overview@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Voice agents@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Live translation@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Transcription@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN + Realtime transcription@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN + Speech to text@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Speech generation@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Realtime prompting guide@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Connection methods@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN + WebRTC@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN + WebSocket@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN + SIP@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Realtime sessions@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN + Managing conversations@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN + Voice activity detection@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN + Realtime with tools@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN + Webhooks and server-side controls@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN + Managing costs@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Specialized models@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Image generation@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Video generation@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Deep research@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Embeddings@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Moderation@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Going live@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Production best practices@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Workload identity federation@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN + Overview@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN + Kubernetes@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN + AWS@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN + Microsoft Azure@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN + Google Cloud@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN + GitHub Actions@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN + SPIFFE@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Deployment checklist@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Amazon Bedrock@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Latency optimization@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN + Overview@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN + Predicted Outputs@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN + Priority processing@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Cost optimization@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN + Overview@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN + Batch@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN + Flex processing@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Accuracy optimization@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Safety@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN + Safety best practices@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN + Safety checks@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN + Cybersecurity checks@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN + Under 18 API Guidance@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Legacy APIs@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Agent Builder@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN + Overview@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN + Migration guide@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN + Node reference@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN + Safety in building agents@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Evals@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN + Getting started@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN + Working with evals@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN + Prompt optimizer@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN + External models@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN + Best practices@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN + Graders@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Fine-tuning@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN + Optimization cycle@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN + Supervised fine-tuning@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN + Vision fine-tuning@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN + Direct preference optimization@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN + Reinforcement fine-tuning@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN + RFT use cases@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN + Best practices@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Assistants API@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN + Migration guide@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN + Deep dive@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN + Tools@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Resources@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Terms and policies@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Changelog@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Your data@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Permissions@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Rate limits@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * IP egress ranges@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Admin APIs@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Deprecations@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * MCP for deep research@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Developer mode@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * ChatGPT Actions@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN + Introduction@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN + Getting started@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN + Actions library@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN + Authentication@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN + Production@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN + Data retrieval@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN + Sending files@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Primary navigation@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN API Codex ChatGPT Docs Use cases Resources Resources@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Search docs@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Suggested@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN response_formatreasoning_effortstreamingtools@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Overview Models Agents Tools Voice & Audio Production API reference@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Overview Models Agents Tools Voice & Audio Production API reference Docs section Overview@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Home@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Get started@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Quickstart@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Using GPT-5.6@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Key concepts@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Core concepts@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Responses API@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Conversation state@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Background mode@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Streaming@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * WebSocket mode@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Multi-agent@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Webhooks@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * File inputs@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Compaction@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Counting tokens@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN SDKs and CLI@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * OpenAI SDK@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * OpenAI CLI@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Resources@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Changelog@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Deprecations@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Supported countries@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * OpenAI Crawlers@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Terms and policies@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Legacy APIs@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Agent Builder@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN + Overview@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN + Migration guide@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN + Node reference@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN + Safety in building agents@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Evals@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN + Getting started@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN + Working with evals@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN + Prompt optimizer@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN + External models@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN + Best practices@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN + Graders@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Fine-tuning@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN + Optimization cycle@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN + Supervised fine-tuning@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN + Vision fine-tuning@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN + Direct preference optimization@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN + Reinforcement fine-tuning@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN + RFT use cases@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN + Best practices@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Assistants API@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN + Migration guide@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN + Deep dive@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN + Tools@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Model catalog@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Choose a model@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Pricing@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Model selection@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Text and code@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Text generation@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Code generation@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Structured output@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Prompting@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Overview@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Prompt engineering@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Citation formatting@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Migration guide@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Prompt generation@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Frontend prompting@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Reasoning@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Reasoning models@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Reasoning best practices@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Images and video@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Images and vision@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Image generation@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Video generation@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Realtime and audio@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Audio and speech@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Overview@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Voice agents@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Specialized models@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Deep research@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Embeddings@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Moderation@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Overview@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Agents SDK@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Quickstart@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Agent definitions@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Models and providers@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Running agents@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Sandbox agents@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Orchestration@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Guardrails@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Results and state@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Integrations and observability@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Evaluate agent workflows@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN ChatKit@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Overview@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Customize@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Widgets@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Actions@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Advanced integrations@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Overview@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Function calling@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Search and retrieval@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Web search@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * File search@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Retrieval@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Connect tools and data@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * MCP and Connectors@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Secure MCP Tunnel@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Build tool workflows@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Skills@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Tool search@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Programmatic tool calling@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Computer and code@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Shell@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Computer use@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Apply Patch@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Local shell@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Code interpreter@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Media@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Image generation@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Overview@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Get started@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Voice agents@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Live translation@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Realtime prompting guide@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Audio@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Audio and speech@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Realtime transcription@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Speech to text@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Speech generation@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Connection methods@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * WebRTC@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * WebSocket@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * SIP@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Sessions and operations@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Managing conversations@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Voice activity detection@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Realtime with tools@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Webhooks and server-side controls@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Managing costs@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Go live@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Production best practices@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Deployment checklist@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Performance and quality@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Latency optimization@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Predicted Outputs@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Priority processing@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Accuracy optimization@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Cost and throughput@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Cost optimization@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Prompt caching@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Batch@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Flex processing@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Safety and governance@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Safety best practices@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Red teaming@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Safety checks@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN + Cybersecurity checks@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN + Under 18 API Guidance@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Your data@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Permissions@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Infrastructure and access@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Private Link@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Workload identity federation@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN + Kubernetes@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN + AWS@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN + Microsoft Azure@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN + Google Cloud@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN + GitHub Actions@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN + SPIFFE@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * IP egress ranges@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Amazon Bedrock@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Operations@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Rate limits@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Admin APIs@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Error codes@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Docs Use cases@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Docs Use cases Docs section Docs@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Apps SDK Workspace Agents Commerce Ads@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Apps SDK Workspace Agents Commerce Ads Docs section Select...@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Home@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Quickstart@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Core Concepts@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * MCP Apps in ChatGPT@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * MCP Server@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * UX principles@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * UI guidelines@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Plan@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Research use cases@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Define tools@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Design components@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Build@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Set up your server@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Build your ChatGPT UI@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Authenticate users@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Manage state@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Monetize your app@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Examples@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Deploy@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Deploy your app@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Connect from ChatGPT@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Test your integration@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Prepare your app@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Conversion apps@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Restaurant reservation spec@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Product checkout spec@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Guides@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Optimize Metadata@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Security & Privacy@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Troubleshooting@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Resources@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Changelog@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * App guidelines@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Reference@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Home@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Get started@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Trigger workspace agent runs@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Authenticate with Workspace Agent access tokens@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Home@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Guides@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Get started@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Best practices@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN File Upload@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Overview@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Products@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN API@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Overview@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Feeds@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Products@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Promotions@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Ads Overview@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Measurement@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * JavaScript Pixel@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Image tag@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Conversions API@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Supported events@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Advertiser API@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Overview@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Quickstart@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Product feeds@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Campaign Targeting@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN API Reference@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Authentication@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Campaigns@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Ad Groups@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Ads@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Ad Account@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Insights@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Files@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Overview Features Configuration Developers Security Administration Use Cases Resources@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Overview Features Configuration Developers Security Administration Use Cases Resources Docs section Overview@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Home@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Get started@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Quickstart@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Use ChatGPT@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Get started with Work@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Import from another agent@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Foundations@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Prompting@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Personalize ChatGPT@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Skills & Plugins@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Permissions@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Explore@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * What's new@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Models@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Pricing@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Glossary@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Available on@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * ChatGPT desktop app@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * ChatGPT on the web@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Codex CLI@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Codex IDE extension@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Codex cloud@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Releases@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Changelog@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Feature Maturity@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Open Source@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Overview@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Workflows@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Projects, chats, and tasks@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Sites@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Visualizations@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Scheduled tasks@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Long-running work@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Notifications@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Pets@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Capabilities@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Browser@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Computer use@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Plugins@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Web search@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Image generation@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Image inputs@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Appshots@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Chrome extension@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Work with files@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Reference@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Commands@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Slash commands@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Settings@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Troubleshooting@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Overview@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Customization@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Overview@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Memories@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Chronicle@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Config file@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Config Basics@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Advanced Config@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Config Reference@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Environment Variables@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Sample Config@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Agent configuration@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * AGENTS.md@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Subagents@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Speed@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Rules@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Extend ChatGPT and Codex@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Record & Replay@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * MCP@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Windows@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Desktop app@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Windows sandbox@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * WSL@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Overview@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Development workflows@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Code review@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Integrated terminal@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Extend and automate@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Build skills@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Build plugins@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Build an app@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Submit plugins@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Hooks@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Environments@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Modes@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Local environments@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Cloud environment@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Git worktrees@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Build with Codex@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Codex SDK@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * App Server@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * MCP Server@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * GitHub Action@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Non-interactive mode@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Third-party integrations@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * GitHub@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Slack@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Linear@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Reference@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * CLI customization@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Developer commands@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Developer settings@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Overview@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Permissions@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Profiles@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Sandboxing@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Auto-review@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Agent approvals & security@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Internet access@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Codex Security@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Overview@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * FAQ@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Codex Security plugin@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN + Quickstart@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN + Run a security scan@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN + Run a deep scan@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN + Review code changes@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN + Triage a backlog@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN + Fix findings@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN + Export and track findings@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN + Changelog@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Codex Security cloud setup@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Improving the threat model@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Safety@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Cyber Safety@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Overview@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Getting started@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Admin rollout guide@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * ChatGPT Work Admin FAQ@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Identity and authentication@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Authentication overview@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Access tokens@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Workspace access, policy, and models@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Groups and provisioning@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Roles and workspace permissions@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Managed configuration@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * HIPAA configuration@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Workspace model availability@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Plugin and connector controls@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Plugin controls@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Skill controls@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Usage, governance, and compliance@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Governance@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Workspace analytics@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Analytics API@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Compliance API and audit events@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Deployment and model providers@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Windows app deployment@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Remote connections@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Amazon Bedrock@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Explore use cases@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Collections@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Home@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Videos@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Showcase@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * OpenAI Academy@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Online trainings@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Community@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Codex Ambassadors@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Codex for Students@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Codex for Open Source@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Meetups@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Blog@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Company blog@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Developer blog@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Explore use cases@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Collections@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Home@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Videos@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Showcase@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * OpenAI Academy@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Online trainings@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Community@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Codex Ambassadors@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Codex for Students@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Codex for Open Source@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Meetups@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Blog@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Company blog@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Developer blog@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Showcase Blog Cookbook Learn Community@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Showcase Blog Cookbook Learn Community Docs section Select...@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * All posts@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Recent@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Mastering remote engineering work from your phone@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Making private MCP servers reachable without making them public@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * How Perplexity Brought Voice Search to Millions Using the Realtime API@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Designing delightful frontends with GPT-5.4@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * From prompts to products: One year of Responses@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Topics@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * General@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * API@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Apps SDK@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Audio@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Codex@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Home@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Topics@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Agents@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Evals@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Multimodal@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Text@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Guardrails@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Optimization@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * ChatGPT@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Codex@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * gpt-oss@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Contribute@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Cookbook on GitHub@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Home@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * OpenAI Developers plugin@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Docs MCP@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Categories@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Demo apps@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Videos@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Topics@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Agents@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Audio & Voice@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Computer Use@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Codex@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Evals@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * gpt-oss@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Fine-tuning@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Image generation@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Scaling@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Tools@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Video generation@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Community@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Programs@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Codex Ambassadors@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Codex for Students@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Codex for Open Source@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * OpenAI for Startups@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Events@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Meetups@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Spaces@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Developer Forum@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Discord@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * Reddit@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN * X@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN API Dashboard@changed_PLACEMARKER_CLOSED\n@changed_PLACEMARKER_OPEN Try ChatGPT@changed_PLACEMARKER_CLOSED\n@changed_into_PLACEMARKER_OPEN * Codex Micro@changed_into_PLACEMARKER_CLOSED", "markup_text_links_to_html_links": false, "restock": {}, "notification_timestamp": 1784178762.40924, "prev_snapshot": " Get started\n\n * Overview\n * Quickstart\n * Models\n * Pricing\n * SDKs and CLI\n + OpenAI SDK\n + Agents SDK\n + OpenAI CLI\n * Latest: GPT-5.5\n * Prompt guidance\n\n Core concepts\n\n * Text generation\n * Code generation\n * Images and vision\n * Audio and speech\n * Structured output\n * Function calling\n * Responses API\n * Using tools\n\n Agents SDK\n\n * Overview\n * Quickstart\n * Agent definitions\n * Models and providers\n * Running agents\n * Sandbox agents\n * Orchestration\n * Guardrails\n * Results and state\n * Integrations and observability\n * Evaluate agent workflows\n * Voice agents\n * ChatKit\n + Overview\n + Customize\n + Widgets\n + Actions\n + Advanced integrations\n\n Tools\n\n * Web search\n * MCP and Connectors\n + Secure MCP Tunnel\n * Skills\n * Shell\n * Computer use\n * File search and retrieval\n + File search\n + Retrieval\n * Tool search\n * More tools\n + Apply Patch\n + Local shell\n + Image generation\n + Code interpreter\n\n Run and scale\n\n * Conversation state\n * Background mode\n * Streaming\n * WebSocket mode\n * Webhooks\n * File inputs\n * Context management\n + Compaction\n + Counting tokens\n + Prompt caching\n * Prompting\n + Overview\n + Prompt engineering\n + Citation formatting\n + Migration guide\n * Reasoning\n + Reasoning models\n + Reasoning best practices\n\n Evaluation\n\n * Red teaming\n\n Realtime and audio\n\n * Overview\n * Voice agents\n * Live translation\n * Transcription\n + Realtime transcription\n + Speech to text\n * Speech generation\n * Realtime prompting guide\n * Connection methods\n + WebRTC\n + WebSocket\n + SIP\n * Realtime sessions\n + Managing conversations\n + Voice activity detection\n + Realtime with tools\n + Webhooks and server-side controls\n + Managing costs\n\n Specialized models\n\n * Image generation\n * Video generation\n * Deep research\n * Embeddings\n * Moderation\n\n Going live\n\n * Production best practices\n * Workload identity federation\n + Overview\n + Kubernetes\n + AWS\n + Microsoft Azure\n + Google Cloud\n + GitHub Actions\n + SPIFFE\n * Deployment checklist\n * Amazon Bedrock\n * Latency optimization\n + Overview\n + Predicted Outputs\n + Priority processing\n * Cost optimization\n + Overview\n + Batch\n + Flex processing\n * Accuracy optimization\n * Safety\n + Safety best practices\n + Safety checks\n + Cybersecurity checks\n + Under 18 API Guidance\n\n Legacy APIs\n\n * Agent Builder\n + Overview\n + Migration guide\n + Node reference\n + Safety in building agents\n * Evals\n + Getting started\n + Working with evals\n + Prompt optimizer\n + External models\n + Best practices\n + Graders\n * Fine-tuning\n + Optimization cycle\n + Supervised fine-tuning\n + Vision fine-tuning\n + Direct preference optimization\n + Reinforcement fine-tuning\n + RFT use cases\n + Best practices\n * Assistants API\n + Migration guide\n + Deep dive\n + Tools\n\n Resources\n\n * Terms and policies\n * Changelog\n * Your data\n * Permissions\n * Rate limits\n * IP egress ranges\n * Admin APIs\n * Deprecations\n * MCP for deep research\n * Developer mode\n * ChatGPT Actions\n + Introduction\n + Getting started\n + Actions library\n + Authentication\n + Production\n + Data retrieval\n + Sending files\n Primary navigation\n API Codex ChatGPT Docs Use cases Resources Resources\n Search docs\n\n Suggested\n\n response_formatreasoning_effortstreamingtools\n Overview Models Agents Tools Voice & Audio Production API reference\n Overview Models Agents Tools Voice & Audio Production API reference Docs section Overview\n * Home\n\n Get started\n\n * Quickstart\n * Using GPT-5.6\n * Key concepts\n\n Core concepts\n\n * Responses API\n * Conversation state\n * Background mode\n * Streaming\n * WebSocket mode\n * Multi-agent\n * Webhooks\n * File inputs\n * Compaction\n * Counting tokens\n\n SDKs and CLI\n\n * OpenAI SDK\n * OpenAI CLI\n\n Resources\n\n * Changelog\n * Deprecations\n * Supported countries\n * OpenAI Crawlers\n * Terms and policies\n\n Legacy APIs\n\n * Agent Builder\n + Overview\n + Migration guide\n + Node reference\n + Safety in building agents\n * Evals\n + Getting started\n + Working with evals\n + Prompt optimizer\n + External models\n + Best practices\n + Graders\n * Fine-tuning\n + Optimization cycle\n + Supervised fine-tuning\n + Vision fine-tuning\n + Direct preference optimization\n + Reinforcement fine-tuning\n + RFT use cases\n + Best practices\n * Assistants API\n + Migration guide\n + Deep dive\n + Tools\n * Model catalog\n\n Choose a model\n\n * Pricing\n * Model selection\n\n Text and code\n\n * Text generation\n * Code generation\n * Structured output\n\n Prompting\n\n * Overview\n * Prompt engineering\n * Citation formatting\n * Migration guide\n * Prompt generation\n * Frontend prompting\n\n Reasoning\n\n * Reasoning models\n * Reasoning best practices\n\n Images and video\n\n * Images and vision\n * Image generation\n * Video generation\n\n Realtime and audio\n\n * Audio and speech\n * Overview\n * Voice agents\n\n Specialized models\n\n * Deep research\n * Embeddings\n * Moderation\n * Overview\n\n Agents SDK\n\n * Quickstart\n * Agent definitions\n * Models and providers\n * Running agents\n * Sandbox agents\n * Orchestration\n * Guardrails\n * Results and state\n * Integrations and observability\n * Evaluate agent workflows\n\n ChatKit\n\n * Overview\n * Customize\n * Widgets\n * Actions\n * Advanced integrations\n * Overview\n * Function calling\n\n Search and retrieval\n\n * Web search\n * File search\n * Retrieval\n\n Connect tools and data\n\n * MCP and Connectors\n * Secure MCP Tunnel\n\n Build tool workflows\n\n * Skills\n * Tool search\n * Programmatic tool calling\n\n Computer and code\n\n * Shell\n * Computer use\n * Apply Patch\n * Local shell\n * Code interpreter\n\n Media\n\n * Image generation\n * Overview\n\n Get started\n\n * Voice agents\n * Live translation\n * Realtime prompting guide\n\n Audio\n\n * Audio and speech\n * Realtime transcription\n * Speech to text\n * Speech generation\n\n Connection methods\n\n * WebRTC\n * WebSocket\n * SIP\n\n Sessions and operations\n\n * Managing conversations\n * Voice activity detection\n * Realtime with tools\n * Webhooks and server-side controls\n * Managing costs\n\n Go live\n\n * Production best practices\n * Deployment checklist\n\n Performance and quality\n\n * Latency optimization\n * Predicted Outputs\n * Priority processing\n * Accuracy optimization\n\n Cost and throughput\n\n * Cost optimization\n * Prompt caching\n * Batch\n * Flex processing\n\n Safety and governance\n\n * Safety best practices\n * Red teaming\n * Safety checks\n + Cybersecurity checks\n + Under 18 API Guidance\n * Your data\n * Permissions\n\n Infrastructure and access\n\n * Private Link\n * Workload identity federation\n + Kubernetes\n + AWS\n + Microsoft Azure\n + Google Cloud\n + GitHub Actions\n + SPIFFE\n * IP egress ranges\n * Amazon Bedrock\n\n Operations\n\n * Rate limits\n * Admin APIs\n * Error codes\n Docs Use cases\n Docs Use cases Docs section Docs\n Apps SDK Workspace Agents Commerce Ads\n Apps SDK Workspace Agents Commerce Ads Docs section Select...\n * Home\n * Quickstart\n\n Core Concepts\n\n * MCP Apps in ChatGPT\n * MCP Server\n * UX principles\n * UI guidelines\n\n Plan\n\n * Research use cases\n * Define tools\n * Design components\n\n Build\n\n * Set up your server\n * Build your ChatGPT UI\n * Authenticate users\n * Manage state\n * Monetize your app\n * Examples\n\n Deploy\n\n * Deploy your app\n * Connect from ChatGPT\n * Test your integration\n * Prepare your app\n\n Conversion apps\n\n * Restaurant reservation spec\n * Product checkout spec\n\n Guides\n\n * Optimize Metadata\n * Security & Privacy\n * Troubleshooting\n\n Resources\n\n * Changelog\n * App guidelines\n * Reference\n * Home\n\n Get started\n\n * Trigger workspace agent runs\n * Authenticate with Workspace Agent access tokens\n * Home\n\n Guides\n\n * Get started\n * Best practices\n\n File Upload\n\n * Overview\n * Products\n\n API\n\n * Overview\n * Feeds\n * Products\n * Promotions\n * Ads Overview\n\n Measurement\n\n * JavaScript Pixel\n * Image tag\n * Conversions API\n * Supported events\n\n Advertiser API\n\n * Overview\n * Quickstart\n * Product feeds\n * Campaign Targeting\n\n API Reference\n\n * Authentication\n * Campaigns\n * Ad Groups\n * Ads\n * Ad Account\n * Insights\n * Files\n Overview Features Configuration Developers Security Administration Use Cases Resources\n Overview Features Configuration Developers Security Administration Use Cases Resources Docs section Overview\n * Home\n\n Get started\n\n * Quickstart\n * Use ChatGPT\n * Get started with Work\n * Import from another agent\n\n Foundations\n\n * Prompting\n * Personalize ChatGPT\n * Skills & Plugins\n * Permissions\n\n Explore\n\n * What's new\n * Models\n * Pricing\n * Glossary\n\n Available on\n\n * ChatGPT desktop app\n * ChatGPT on the web\n * Codex CLI\n * Codex IDE extension\n * Codex cloud\n\n Releases\n\n * Changelog\n * Feature Maturity\n * Open Source\n * Overview\n\n Workflows\n\n * Projects, chats, and tasks\n * Sites\n * Visualizations\n * Scheduled tasks\n * Long-running work\n * Notifications\n * Pets\n\n Capabilities\n\n * Browser\n * Computer use\n * Plugins\n * Web search\n * Image generation\n * Image inputs\n * Appshots\n * Chrome extension\n * Work with files\n\n Reference\n\n * Commands\n * Slash commands\n * Settings\n * Troubleshooting\n * Overview\n\n Customization\n\n * Overview\n * Memories\n * Chronicle\n\n Config file\n\n * Config Basics\n * Advanced Config\n * Config Reference\n * Environment Variables\n * Sample Config\n\n Agent configuration\n\n * AGENTS.md\n * Subagents\n * Speed\n * Rules\n\n Extend ChatGPT and Codex\n\n * Record & Replay\n * MCP\n\n Windows\n\n * Desktop app\n * Windows sandbox\n * WSL\n * Overview\n\n Development workflows\n\n * Code review\n * Integrated terminal\n\n Extend and automate\n\n * Build skills\n * Build plugins\n * Build an app\n * Submit plugins\n * Hooks\n\n Environments\n\n * Modes\n * Local environments\n * Cloud environment\n * Git worktrees\n\n Build with Codex\n\n * Codex SDK\n * App Server\n * MCP Server\n * GitHub Action\n * Non-interactive mode\n\n Third-party integrations\n\n * GitHub\n * Slack\n * Linear\n\n Reference\n\n * CLI customization\n * Developer commands\n * Developer settings\n * Overview\n\n Permissions\n\n * Profiles\n * Sandboxing\n * Auto-review\n * Agent approvals & security\n * Internet access\n\n Codex Security\n\n * Overview\n * FAQ\n * Codex Security plugin\n + Quickstart\n + Run a security scan\n + Run a deep scan\n + Review code changes\n + Triage a backlog\n + Fix findings\n + Export and track findings\n + Changelog\n * Codex Security cloud setup\n * Improving the threat model\n\n Safety\n\n * Cyber Safety\n * Overview\n\n Getting started\n\n * Admin rollout guide\n * ChatGPT Work Admin FAQ\n\n Identity and authentication\n\n * Authentication overview\n * Access tokens\n\n Workspace access, policy, and models\n\n * Groups and provisioning\n * Roles and workspace permissions\n * Managed configuration\n * HIPAA configuration\n * Workspace model availability\n\n Plugin and connector controls\n\n * Plugin controls\n * Skill controls\n\n Usage, governance, and compliance\n\n * Governance\n * Workspace analytics\n * Analytics API\n * Compliance API and audit events\n\n Deployment and model providers\n\n * Windows app deployment\n * Remote connections\n * Amazon Bedrock\n * Explore use cases\n * Collections\n * Home\n * Videos\n * Showcase\n * OpenAI Academy\n * Online trainings\n\n Community\n\n * Codex Ambassadors\n * Codex for Students\n * Codex for Open Source\n * Meetups\n\n Blog\n\n * Company blog\n * Developer blog\n * Explore use cases\n * Collections\n * Home\n * Videos\n * Showcase\n * OpenAI Academy\n * Online trainings\n\n Community\n\n * Codex Ambassadors\n * Codex for Students\n * Codex for Open Source\n * Meetups\n\n Blog\n\n * Company blog\n * Developer blog\n Showcase Blog Cookbook Learn Community\n Showcase Blog Cookbook Learn Community Docs section Select...\n * All posts\n\n Recent\n\n * Mastering remote engineering work from your phone\n * Making private MCP servers reachable without making them public\n * How Perplexity Brought Voice Search to Millions Using the Realtime API\n * Designing delightful frontends with GPT-5.4\n * From prompts to products: One year of Responses\n\n Topics\n\n * General\n * API\n * Apps SDK\n * Audio\n * Codex\n * Home\n\n Topics\n\n * Agents\n * Evals\n * Multimodal\n * Text\n * Guardrails\n * Optimization\n * ChatGPT\n * Codex\n * gpt-oss\n\n Contribute\n\n * Cookbook on GitHub\n * Home\n * OpenAI Developers plugin\n * Docs MCP\n\n Categories\n\n * Demo apps\n * Videos\n\n Topics\n\n * Agents\n * Audio & Voice\n * Computer Use\n * Codex\n * Evals\n * gpt-oss\n * Fine-tuning\n * Image generation\n * Scaling\n * Tools\n * Video generation\n * Community\n\n Programs\n\n * Codex Ambassadors\n * Codex for Students\n * Codex for Open Source\n * OpenAI for Startups\n\n Events\n\n * Meetups\n\n Spaces\n\n * Developer Forum\n * Discord\n * Reddit\n * X\n API Dashboard\n Try ChatGPT", "preview_url": "https://changedetection.benfou.fr/preview/40d1d040-eab5-44c8-a830-96ce77c471dd", "screenshot": null, "timestamp_from": null, "timestamp_to": null, "triggered_text": "", "llm_summary": null, "llm_intent": null, "uuid": "40d1d040-eab5-44c8-a830-96ce77c471dd", "watch_mime_type": "text/html; charset=utf-8", "watch_tag": "IA", "watch_title": "API OpenIA", "watch_url": "https://developers.openai.com/api/docs/changelog", "watch_uuid": "40d1d040-eab5-44c8-a830-96ce77c471dd", "notification_urls": ["jsons://n8n.benfou.fr/webhook/c8bb38ef-b9d8-45d0-837d-c239411db01b/?title={{ watch_title|urlencode }}"], "notification_title": "Changement d\u00e9tect\u00e9 - {{watch_title}}", "notification_body": "{{watch_url}} a chang\u00e9e.\r\n---\r\n{{diff}}\r\n---\r\n", "notification_format": "htmlcolor", "_llm_result": null, "_llm_intent": "", "_llm_change_summary": "", "edit_url": "https://changedetection.benfou.fr/edit/40d1d040-eab5-44c8-a830-96ce77c471dd"}}]