Code written for Claude Sonnet 5 can break on Claude Sonnet 5.5 in five ways
Code written for Claude Sonnet 5 can break on Claude Sonnet 5.5 in five ways. To turn off up-front thinking, send thinking: {"type": "between_tools"} instead of "disabled", at high effort or below. Forced tool use (tool_choice types any and tool) returns a 400 error. Thinking blocks are tied to the model and the conversation.
Also shipped Sep 28, 2026Anthropic in September 2026 →
Manage and de-risk your API keys with new Security Center
OpenRouter introduced a Security Center to help users manage and audit API keys. The tool identifies risky keys and allows users to lock access to specific networks.
Session replay now has its own billing category
New Relic now reports session replay data (both Browser and Mobile) under its own billing category in Data ingestion and usage, instead of under browser events and mobile events. This is a reporting change only: your session replay functionality, pricing, and total data ingested are unaffected.
Create custom URLs , like example.cloud.run , that are easy to remember and globally available for your Cloud...
Create custom URLs , like example.cloud.run , that are easy to remember and globally available for your Cloud Run services. No DNS configuration, load balancers, or extra costs required ( Preview ).
To improve security, removed the cloudsql.instances.export permission from the following roles: - Cloud SQL...
To improve security, removed the cloudsql.instances.export permission from the following roles: Cloud SQL Viewer ( roles/cloudsql.viewer ) Basic Reader ( roles/reader ) Basic Viewer (Legacy) ( roles/viewer ) To retain export capabilities, assign the Cloud SQL Editor ( roles/cloudsql.editor ) role or update custom roles to include the...
Prompt and preview in one view with Agent Runners
Agent Runners now shows you a preview right next to your prompt. Previously, you’d get a link to the Deploy Preview and have to open it in a separate tab to see what the agent changed. Now the preview appears inline, so you can review the result and write your next prompt without switching back and forth.
More from Claude PlatformEvery Claude Platform release →
We've launched Claude Sonnet 5.5 (claude-sonnet-5-5)
We've launched Claude Sonnet 5.5 (claude-sonnet-5-5). It's available on the Claude API, Claude in Amazon Bedrock, Claude Platform on AWS, Claude on Google Cloud, and Claude in Microsoft Foundry. For its context window, output limits, and prices, see the Claude Sonnet 5.5 model page.
Thinking blocks that Claude Sonnet 5.5 produces work only in the account that produced them, or in an account
Thinking blocks that Claude Sonnet 5.5 produces work only in the account that produced them, or in an account linked to it. When another account sends one of these blocks, the API drops the block before the model sees it, and the request succeeds. Blocks from earlier models aren't affected. See Preserved thinking.
We announced the deprecation of the Claude Sonnet 4.5 model (claude-sonnet-4-5-20250929), with retirement on
We announced the deprecation of the Claude Sonnet 4.5 model (claude-sonnet-4-5-20250929), with retirement on the Claude API scheduled for November 30, 2026. We recommend migrating to Claude Sonnet 5.5. Read more in Model deprecations.
In Python SDK 1.10.0, TypeScript SDK 0.130.0, C# SDK 12.52.0, Go SDK 1.77.0, Java SDK 2.67.0, PHP SDK 0.53.0
In Python SDK 1.10.0, TypeScript SDK 0.130.0, C# SDK 12.52.0, Go SDK 1.77.0, Java SDK 2.67.0, PHP SDK 0.53.0, Ruby SDK 1.75.0, and version 1.37.0 of the ant CLI, the Admin API methods for organization info, members, invites, workspaces, API keys, rate limits, service accounts, workload identity federation, customer-managed encryption keys, and compliance...
We've added a line field to the Models API
We've 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. Claude Opus 4.5 and Claude Opus 4.6 both report opus, for example. Use line to group models without parsing their IDs. line is null for a model that belongs to no line. See Using the Models API.
Build plugins for Claude
Developers can now submit plugins to the Claude directory via a new portal. The portal includes features to track submissions through review and view usage analytics.
We're expanding which refusals are billed to include refusals that arrive before any output when
We're expanding which refusals are billed to include refusals that arrive before any output when stop_details.category is "bio", "frontier_llm", or "reasoning_extraction", the categories where we measure low volumes of false positives. Mid-stream refusals were already billed.