---
name: Claude Platform
slug: claude-platform-releases
type: scrape
source_url: https://platform.claude.com/docs/en/release-notes/overview
organization: Anthropic
organization_slug: anthropic
total_releases: 598
latest_date: 2026-10-08
last_updated: 2026-10-06
tracking_since: 2024-05-10
canonical: https://releases.sh/anthropic/claude-platform-releases
organization_url: https://releases.sh/anthropic
---

<Summary type="rolling" window-days="90" release-count="15">
Claude Platform shipped a major model refresh and expanded its agentic capabilities. Opus 4.7 arrived with improved reasoning and broke API compatibility, requiring migration from earlier versions; Sonnet 4.6 landed as a faster balanced option with extended thinking and 1M context now generally available at standard pricing. The platform grew its agent-building toolkit with Managed Agents (fully sandboxed autonomous operation), an advisor tool (pairing fast executors with high-intelligence guidance), and the `ant` CLI for streamlined local development. Automatic caching simplified request optimization, the compaction API enabled effectively infinite conversations through server-side summarization, and capability discovery moved into the Models API itself.
</Summary>

<Summary type="monthly" period="March 2026" release-count="4">
Extended output and expanded context reached general availability this month. The 1M token context window graduated to standard pricing for Claude Opus 4.6 and Sonnet 4.6, with the media limit per request raised to 600 images or PDFs. The max_tokens cap on Batches API grew to 300k for longer single-turn outputs, while a new display field for extended thinking lets developers omit thinking blocks for faster streaming. The Models API gained capability fields exposing max_input_tokens, max_tokens, and a capabilities object for runtime model discovery. The 1M context beta for Sonnet 4.5 and 4 retires April 30, after which requests exceeding 200k tokens will error on those models.
</Summary>

<Release date="October 8, 2026" published="2026-10-08T00:00:00.000Z" url="https://platform.claude.com/docs/en/release-notes/overview#compliance-api-now-returns-chats-from-unified-claude-experie">
## Compliance API now returns chats from unified Claude experience

The Compliance API chat endpoints now also return chats from the unified Claude experience, in beta for Claude Enterprise organizations, with your existing Compliance Access Key.
</Release>

<Release date="October 7, 2026" published="2026-10-07T00:00:00.000Z" url="https://platform.claude.com/docs/en/release-notes/overview#claude-managed-agents-web-fetch-now-restricted-to-urls-in-pr">
## Claude Managed Agents: web_fetch now restricted to URLs in prior context

The `web_fetch` tool in Claude Managed Agents now fetches only URLs that have already appeared in the session (in user messages, web_search results, or previously fetched pages). URLs appearing only in Claude's output, system prompt, attached documents, or tool outputs return `url_not_in_prior_context` error. This reduces data exfiltration risk.
</Release>

<Release date="October 7, 2026" published="2026-10-07T00:00:00.000Z" url="https://platform.claude.com/docs/en/release-notes/overview#claude-managed-agents-web-tools-now-respect-allowed-hosts-in">
## Claude Managed Agents: web tools now respect allowed_hosts in limited networking

In Claude Managed Agents, cloud environments with `limited` networking now apply `allowed_hosts` to `web_search` and `web_fetch` tools. `web_fetch` calls for URLs on hosts not matching `allowed_hosts` return `url_not_allowed` error. `web_search` omits results from such hosts. When `allowed_hosts` lists no hosts, neither tool returns results.
</Release>

<Release date="October 7, 2026" published="2026-10-07T00:00:00.000Z" url="https://platform.claude.com/docs/en/release-notes/overview#prompt-cache-reads-pricing-reduced-for-claude-sonnet-5-5">
## Prompt cache reads pricing reduced for Claude Sonnet 5.5

Lowered the price of prompt cache reads on Claude Sonnet 5.5 from $0.20 USD to $0.10 USD per million tokens (0.05x the base input price instead of 0.1x). Cache writes and all other prices unchanged.
</Release>

<Release date="October 7, 2026" published="2026-10-07T00:00:00.000Z" url="https://platform.claude.com/docs/en/release-notes/overview#claude-haiku-5-5-launched-with-1m-token-context-and-adaptive">
## Claude Haiku 5.5 launched with 1M token context and adaptive thinking

Launched Claude Haiku 5.5 (`claude-haiku-5-5`), tuned for high-volume and latency-sensitive work. Features include 1M token context window, 128k max output tokens, and adaptive thinking with effort parameter. Available on Claude API, Amazon Bedrock, Claude Platform on AWS, Google Cloud, and Microsoft Foundry.
</Release>

<Release date="October 7, 2026" published="2026-10-07T00:00:00.000Z" url="https://platform.claude.com/docs/en/release-notes/overview#browser-and-computer-use-sdk-toolsets-now-available-in-beta">
## Browser and computer use SDK toolsets now available in beta

Python and TypeScript SDKs now include classes in beta for the browser use tool and computer use tool. Subclass one and write one method per tool against your own browser or desktop automation. SDK runs the tool loop, URL and file policies, and approval callback.
</Release>

<Release date="October 7, 2026" published="2026-10-07T00:00:00.000Z" url="https://platform.claude.com/docs/en/release-notes/overview#claude-max-and-team-plans-now-include-monthly-api-credits">
## Claude Max and Team plans now include monthly API credits

Claude Max and Team plans now include monthly API credits. Users can claim them through the API credits for subscribers feature.
</Release>

<Release date="October 6, 2026" published="2026-10-06T00:00:00.000Z" url="https://platform.claude.com/docs/en/release-notes/overview#models-api-now-reports-server-tools-capabilities">
## Models API now reports server tools capabilities

Added `capabilities.server_tools` to the Models API. `GET /v1/models` and `GET /v1/models/{model_id}` now report whether each model accepts web search and code execution tools. `capabilities.server_tools.code_execution` indicates code execution tool support.
</Release>

<Release date="October 5, 2026" published="2026-10-05T00:00:00.000Z" url="https://platform.claude.com/docs/en/release-notes/overview#models-api-capabilities-thinking-types-disabled">
## Models API: capabilities.thinking.types.disabled

Added `capabilities.thinking.types.disabled` to the Models API. `GET /v1/models` and `GET /v1/models/{model_id}` now report whether each model accepts `thinking: {type: "disabled"}`, which turns thinking off.
</Release>

<Release date="October 1, 2026" published="2026-10-01T00:00:00.000Z" url="https://platform.claude.com/docs/en/release-notes/overview#models-api-model-line-field-added">
## Models API: model line field added

Added a `line` field to the Models API. `GET /v1/models` and `GET /v1/models/{model_id}` now return the model line each model belongs to (e.g., Claude Opus 4.5 and Claude Opus 4.6 both report `opus`). Use `line` to group models without parsing their IDs. `line` is `null` for models that belong to no line.
</Release>

<Release date="September 30, 2026" published="2026-09-30T00:00:00.000Z" url="https://platform.claude.com/docs/en/release-notes/overview#claude-sonnet-4-5-deprecation-announced">
## Claude Sonnet 4.5 deprecation announced

Claude Sonnet 4.5 (`claude-sonnet-4-5-20250929`) is deprecated with retirement scheduled for November 30, 2026. Users are recommended to migrate to Claude Sonnet 5.5. See [Model deprecations](https://platform.claude.com/docs/en/about-claude/model-deprecations).
</Release>

<Release date="September 28, 2026" published="2026-09-28T00:00:00.000Z" url="https://platform.claude.com/docs/en/release-notes/overview#claude-sonnet-5-5-launched">
## Claude Sonnet 5.5 launched

Claude Sonnet 5.5 (`claude-sonnet-5-5`) is now available on the Claude API, Amazon Bedrock, AWS, Google Cloud, and Microsoft Foundry. The model supports a 1M token context window and 128k max output tokens. Code written for Claude Sonnet 5 may break on Claude Sonnet 5.5 in five ways: turning off up-front thinking requires `thinking: {"type": "between_tools"}` instead of `"disabled"` at effort `high` or below; forced tool use (`tool_choice` types `any` and `tool`) returns a 400 error; thinking blocks are tied to the model and conversation; the earlier `computer_20251124` computer use tool isn't accepted on Claude API and Google Cloud; the advisor tool rejects Claude Opus 4.8, 4.7, and Sonnet 5 as advisors. Thinking blocks produced by Sonnet 5.5 work only in the account that produced them or linked accounts. See [What's new in Claude Sonnet 5.5](https://platform.claude.com/docs/en/models/sonnet-5-5/whats-new-sonnet-5-5) and [migration guide](https://platform.claude.com/docs/en/models/sonnet-5-5/migration-guide).
</Release>

<Release date="September 24, 2026" published="2026-09-24T00:00:00.000Z" url="https://platform.claude.com/docs/en/release-notes/overview#compliance-api-updates-and-refusal-billing-changes">
## Compliance API updates and refusal billing changes

Resuming billing for refusals that arrive before any output when `stop_details.category` is `"bio"`, `"frontier_llm"`, or `"reasoning_extraction"`. The Compliance API local session endpoints are out of beta for Claude for Microsoft 365 sessions in Excel, PowerPoint, Word, and Outlook. The Compliance API Activity Feed no longer returns file names, project document names, or artifact titles.
</Release>

<Release date="September 24, 2026" published="2026-09-24T00:00:00.000Z" url="https://platform.claude.com/docs/en/release-notes/overview#refusal-billing-changes-and-compliance-api-updates">
## Refusal billing changes and Compliance API updates

Billing for refusals before any output resumes for `stop_details.category` values `"bio"`, `"frontier_llm"`, and `"reasoning_extraction"`. Mid-stream refusals were already billed. Compliance API local session endpoints are out of beta for Claude for Microsoft 365 sessions in Excel, PowerPoint, Word, and Outlook. Activity Feed no longer returns file names, project document names, or artifact titles; `filename` and `title` fields are now empty or omitted. Use a Compliance Access Key with `read:compliance_user_data` scope to look up names by ID.
</Release>

<Release date="September 23, 2026" published="2026-09-23T00:00:00.000Z" url="https://platform.claude.com/docs/en/release-notes/overview#cache-diagnostics-out-of-beta">
## Cache diagnostics out of beta

Cache diagnostics is out of beta on the Claude API and no longer requires the `cache-diagnosis-2026-04-07` beta header. Include the `diagnostics` object on a Messages request to opt in. Responses from `POST /v1/messages` now always include the `diagnostics` field, which is `null` when the request did not include the `diagnostics` object.
</Release>

<Release date="September 22, 2026" published="2026-09-22T00:00:00.000Z" url="https://platform.claude.com/docs/en/release-notes/overview#claude-opus-5-5-launched">
## Claude Opus 5.5 launched

Claude Opus 5.5 (`claude-opus-5-5`) is available for long-running agentic coding and knowledge work with a 1M token context window by default, 128k max output tokens, and always-on adaptive thinking at $4 / $20 USD per MTok (vs $5 / $25 for Opus 5). Available on Claude API, Amazon Bedrock, AWS, Google Cloud, and Microsoft Foundry. Thinking cannot be disabled: `thinking: {"type": "disabled"}` and `thinking: {"type": "enabled", ...}` return 400 errors. `tool_choice` types `any` and `tool` also return 400 errors; use `auto` with strict tool use. Computer use requires `computer_toolset_20260801` on Claude API and Google Cloud; earlier `computer_20251124` tool returns 400 error. Fast mode (research preview) is available. Tools can now be defined inside mid-conversation system messages (beta, `inline-tools-2026-09-15` header). See [What's new in Claude Opus 5.5](https://platform.claude.com/docs/en/models/opus-5-5/whats-new-opus-5-5) and [migration guide](https://platform.claude.com/docs/en/models/opus-5-5/migration-guide).
</Release>

<Release date="September 22, 2026" published="2026-09-22T00:00:00.000Z" url="https://platform.claude.com/docs/en/release-notes/overview#mid-conversation-tools-inline-tool-definitions-and-mcp-tools">
## Mid-conversation tools: inline tool definitions and MCP toolsets

Tools can now be defined inside a [mid-conversation system message](https://platform.claude.com/docs/en/build-with-claude/mid-conversation-system-messages#define-tools-in-a-message-beta), in beta on the Claude API with the `inline-tools-2026-09-15` beta header. A `tool_addition` block can carry the tool's full definition (`tool: {"type": "tool_definition", "definition": {...}}`), allowing you to add a tool, change its schema, or move a server tool to a newer version without editing `tools` or invalidating the prompt cache. The same header covers adding and removing tools by reference. With the MCP connector's `mcp-client-2026-09-15` beta header, the definition can be an MCP toolset, and a response records each server's fetched tool list in an `mcp_tool_listing` block.
</Release>

<Release date="September 18, 2026" published="2026-09-18T00:00:00.000Z" url="https://platform.claude.com/docs/en/release-notes/overview#compliance-api-local-sessions-now-include-claude-in-chrome">
## Compliance API local sessions now include Claude in Chrome

The Compliance API local session endpoints now return transcripts of Claude in Chrome sessions (`product_surface` value `claude_in_chrome`), in beta for Claude Enterprise organizations, with existing Compliance Access Key and `read:compliance_user_data` scope.
</Release>

<Release date="September 18, 2026" published="2026-09-18T00:00:00.000Z" url="https://platform.claude.com/docs/en/release-notes/overview#cache-diagnostics-response-field-always-included">
## Cache diagnostics response field always included

For cache diagnostics, responses to requests sending the `cache-diagnosis-2026-04-07` beta header now always include the `diagnostics` field, which is `null` when the request did not include the `diagnostics` object. Previously the field was omitted in that case.
</Release>

<Release date="September 18, 2026" published="2026-09-18T00:00:00.000Z" url="https://platform.claude.com/docs/en/release-notes/overview#compliance-api-claude-in-chrome-sessions-support">
## Compliance API: Claude in Chrome sessions support

The Compliance API local session endpoints now return transcripts of Claude in Chrome sessions (`product_surface` value `claude_in_chrome`), in beta for Claude Enterprise organizations. Uses existing Compliance Access Key and `read:compliance_user_data` scope.
</Release>

<Pagination cursor="2026-09-18T00:00:00.000Z|2026-09-22T13:51:54.181Z|rel_cTZffEnyGLpDpQHKP0Pv5" next="https://releases.sh/anthropic/claude-platform-releases.md?cursor=2026-09-18T00%3A00%3A00.000Z%7C2026-09-22T13%3A51%3A54.181Z%7Crel_cTZffEnyGLpDpQHKP0Pv5&limit=20" />