Mistral Large 4 now available on AI Gateway
By Steven Van ·
The public preview supports text and image input, with a 1M-token context window and access through one API key.
Developers can access Mistral Large 4 through Vercel AI Gateway with one API key, without creating a separate Mistral account. The open-weight model is available in public preview and combines reasoning with text and image understanding.
Text, images and reasoning
Mistral Large 4 has a 1M-token context window shared by prompts and responses, and supports up to 262,144 output tokens, including reasoning tokens. Images can be sent alongside text as message parts and count toward input-token usage. Mistral lists cyber defense, manufacturing, finance and visual grounding among the model's strengths.
Use it in apps and coding agents
Use the model ID mistral/mistral-large-4 with the AI SDK's generateText or streamText functions. OpenAI-compatible Chat Completions, OpenAI Responses and Anthropic Messages APIs are also supported by changing the base URL. AI Gateway authenticates requests and routes them to an available provider.
For coding agents, run npx vercel ai-gateway setup, then select mistral/mistral-large-4 in the agent's model settings. Developers can use an AI Gateway API key or bring their own provider key, and test the model in the playground.
Preview pricing
The model page lists pricing at $0.68 per million input tokens and $2.09 per million output tokens, marked as 50% off.