Mistral opens Responses API in public preview

Mistral AI launched its Responses API in public preview on La Plateforme: the /v1/responses endpoint is compatible with OpenAI's format, so agent pipelines move over by changing the base URL and API key. It works with every Mistral chat model, including the Large 4 flagship, and supports text, image and file inputs, streaming and function calling.

• The /v1/responses endpoint runs on api.mistral.ai across all Mistral chat models
• Output is a sequence of typed items: reasoning, tool calls and messages
• The API is stateless; clients resend prior context with each request
• Large 4 is priced at $1.36 per 1M input and $4.18 per 1M output tokens