What changed: The GET /v1/models and GET /v1/models/{model_id} endpoints of the Models API gained a capabilities.server_tools field reporting whether each model accepts the web search and code execution tools. Code execution support is checked via capabilities.server_tools.code_execution, while whether code can call the request's other tools is reported via the top-level capabilities.code_execution field.
Who is affected
Developers and businesses using the Claude API (Models API) in their integrations.
What it means for your business
For businesses using the Claude API, this means they can now automatically read from the API response whether a given model supports web search and code execution before choosing it.
What to do
Update your integration's model-selection logic to read the capabilities.server_tools field from the GET /v1/models response; check code execution support via capabilities.server_tools.code_execution and whether code can call other tools via the top-level capabilities.code_execution field.
When it takes effect
October 6, 2026
Source: https://platform.claude.com/docs/en/release-notes/overview#october-6-2026
