Skip to main content

We've added capabilities.thinking.types.disabled to the Models API

By Steven Van ·

Both model endpoints report support for turning thinking off, though an incompatible effort level can still cause a request to fail.

Developers can now check whether a model accepts the setting that turns thinking off through Anthropic's Models API. Both GET /v1/models and GET /v1/models/{model_id} return capabilities.thinking.types.disabled.

Reading the support flag

As described in the Models API documentation, supported is false when a model rejects thinking: {type: "disabled"} with a 400 error. It is true for a model that does not support thinking.

A true value does not guarantee that every request with thinking disabled will succeed. The API can still reject the request for another reason, such as an effort level the model does not allow with thinking off.

Anthropic
Anthropic
Anthropic is an AI safety and research company that's working to build reliable, interpretable, and steerable AI systems.
View Anthropic →

Original source

Read We've added capabilities.thinking.types.disabled to the Models API on Creators Toolbox