Releases
Mistral AI opens public preview for 1.05T-parameter Large 4
Mistral AI opened a public preview of Mistral Large 4 on October 6, 2026. The open-weight multimodal model uses a granular Mixture-of-Experts design with 52 billion active parameters, 1.05 trillion in total, and a 1.6B vision encoder. Its context window reaches 1M tokens. API prices are $1.36 per input million tokens and $4.18 per output million.
Mistral AI opened a public preview of Mistral Large 4 on October 6, 2026. The documentation tags the model mistral-large-4 and marks the release Openv26.10. It is an open-weight, general-purpose multimodal system built on a granular Mixture-of-Experts design, activating 52B of its 1.05T total parameters per request. A separate 1.6B-parameter encoder handles vision input.
The context window is 1M tokens. API pricing is $1.36 per million input tokens, $0.14 per million cached input tokens, and $4.18 per million output tokens.
The documented capabilities are Structured Outputs, Function Calling, Document QnA, Prefix, Chat Completions, Batching, Agents & Conversations, and Built-In Tools. The listed routes are /v1/chat/completions, /v1/conversations, /v1/batch, and /v1/agents. Structured Outputs, Function Calling, Document QnA, Prefix, and Chat Completions use /v1/chat/completions. Agents & Conversations and Built-In Tools use /v1/agents and /v1/conversations.
Readers’ Forum
No contributions yet — open the debate.
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.