Claude Platform adds capability discovery API, extended thinking display control
Models API Enhancements
Anthropic has expanded the Models API with capability discovery fields. The GET /v1/models and GET /v1/models/{model_id} endpoints now return max_input_tokens, max_tokens, and a capabilities object, allowing developers to programmatically discover what each model supports without manual documentation lookup.
Extended Thinking Display Control
A new display field for extended thinking gives developers finer control over response streaming. By setting thinking.display: "omitted", developers can receive thinking blocks with empty thinking content while preserving the signature field needed for multi-turn conversations. This enables faster streaming performance without sacrificing conversation continuity, and billing remains unchanged.
Developer Actions
- Update API integrations to leverage new capability fields for dynamic model selection and feature detection
- Optimize extended thinking workflows by using the
display: "omitted"setting for performance-critical applications where thinking content isn't needed in the response - Review the Controlling thinking display documentation for implementation details