Claude Platform Updates: New Models, Deprecations, and Tool Enhancements
Anthropic has released new flagship models, Claude Fable 5 and Claude Opus 4.8, featuring larger context windows and enhanced adaptive thinking. The company is also retiring older models like Claude Sonnet 4 and Opus 4.1, recommending upgrades. Several tools and platform features, including code execution, web search, and managed agents, have received updates and expanded availability on AWS, impacting developers and researchers using the Claude API.
- →Claude Sonnet 4 and Opus 4 models retired
- →Code execution tool updated with time limits
- →Web search tool supports response inclusion parameter
- →Self-hosted sandbox work endpoint available on AWS
- →New models Claude Fable 5 and Claude Mythos 5 launched
Deprecations (1) ›
- Claude Sonnet 4 and Opus 4 models retired
The Claude Sonnet 4 (claude-sonnet-4-20250514) and Claude Opus 4 (claude-opus-4-20250514) models have been retired. Requests to these models will now result in an error, and users are advised to upgrade to Claude Sonnet 4.6 and Claude Opus 4.8 respectively.
Features (4) ›
- Code execution tool updated with time limits
The code execution tool now supports version code_execution_20260521. This version explicitly discloses a 90-second per-cell execution time limit in its description, allowing Claude to better budget for long-running cells.
- Web search tool supports response inclusion parameter
The web search tool now supports the `response_inclusion` parameter. This parameter allows for dropping consumed result blocks from the API response, which is beneficial for agentic workflows.
- Self-hosted sandbox work endpoint available on AWS
The GET /v1/environments/{id}/work endpoint, which lists pending work for self-hosted sandboxes, is now available on Claude Platform on AWS. Authorization is managed by the GetEnvironment IAM action.
- New models Claude Fable 5 and Claude Mythos 5 launched
Anthropic launched Claude Fable 5, their most capable widely released model, and Claude Mythos 5 for Project Glasswing participants. Both support a 1M token context window, 128k max output tokens, and always-on adaptive thinking.
Enhancements (1) ›
- Claude Fable 5 and Mythos 5 use new tokenizer
Claude Fable 5 and Claude Mythos 5 utilize a new tokenizer, resulting in approximately 30% more tokens for the same text compared to models before Claude Opus 4.7. A token counting API is available to measure prompts under this new tokenizer.
https://platform.claude.com/docs/en/release-notes/overview