Mistral AI dated the model release October 6, 2026. The documentation marks it as a Public Preview under Open v26.10. Its model identifier is mistral-large-4. Mistral positions it as a general-purpose multimodal model with open weights and a granular Mixture-of-Experts architecture.
The model has 49B active parameters out of 1.05T total parameters. Its vision encoder contains 1.6B parameters. The context window reaches 1M tokens.
The API supports Structured Outputs, Function Calling, Document QnA, Prefix, and Chat Completions through /v1/chat/completions and /v1/conversations. Batching uses /v1/batch. Agents & Conversations and Built-In Tools use /v1/agents and /v1/conversations.
The page lists prices of $1.36 per million input tokens, $0.14 for cached input, and $4.18 for output. It also displays a second set of figures: $0.68, $0.07, and $2.09 for the same three categories.
Comments
No comments yet — be the first.
Open the discussion
No account or password needed — just enter your e-mail and we’ll send you a one-time sign-in link. First time here? You’re set up automatically.
Your rating will be applied automatically after you sign in.
Check your inbox
We’ve sent a sign-in link to …. Open it on this device — this tab will sign you in automatically.
Nothing arrived? Check your spam folder — and mark the mail as "Not spam" so it lands in your inbox next time.