PyyanThe AI Index

Pyyan / Language Models / Mistral Large 4

Mistral Large 4

A trillion parameters in preview, with the weights promised but not shipped.

previewMistral ↗Verified 2026-10-10

$1.36

per M input

Specifications

Context1Msource · as of 2026-10-10
Input ($/Mtok)$1.36source · as of 2026-10-10
Output ($/Mtok)Not publishedsource · as of 2026-10-10
LicenceNot publishedsource · as of 2026-10-10
Max outputNot publishedsource · as of 2026-10-10
Parameters1.05T total, 49B or 52B activesource · as of 2026-10-10 · low confidence
Released6 Oct 2026source · as of 2026-10-10
Model IDmistral-large-4source · as of 2026-10-10 · medium confidence

Official links

Mistral Large 4 model pageMistral official site

What matters

  • Preview API only: weights are promised by the end of October and the licence is unpublished
  • Sources disagree on active parameters, 49B in early coverage and 52B in the documentation
  • 1M token context, with a 1.6B parameter vision encoder

Best for trying a European frontier model.

Compare with

All 5 side by side →

Mistral Large 4 vs GPT-6 Astra →Mistral Large 4 vs Claude Fable 5.1 →Mistral Large 4 vs Claude Fable 5 →Mistral Large 4 vs Claude Opus 5.5 →Mistral Large 4 vs GPT-6.1 Sol →

Also in Language Models

GPT-6 Astra · OpenAIClaude Fable 5.1 · AnthropicClaude Fable 5 · AnthropicClaude Opus 5.5 · AnthropicGPT-6.1 Sol · OpenAI