Mistral Large 4
Not scored yet.

MistralOct 6, 2026

Capabilities Index

Not run yet

–

Context

Takes text, image

1.05M

Input

Per 1M tokens

$0.68

Output

Per 1M tokens

$2.09

Benchmarks

No independent scores yet.

Epoch AI hasn't run Mistral Large 4 yet. Its scores show up here the morning after they do.

On the index

The top of the index.

#1Claude Opus 5.5Anthropic · $4.00 in167.3
#2GPT-6 AstraOpenAI · $10 in166.4
#3GPT-6.1 SolOpenAI · $2.00 in166.1
#4Claude Sonnet 5.5Anthropic · $2.00 in165.0
#5Claude Fable 5.1Anthropic · $10 in164.7
#6Claude Opus 5Anthropic · $5.00 in162.8

In the news

1 story.

ModelMistral AIOctober 6, 2026

Mistral Large 4 is out, and the weights are open

Mistral's biggest model yet, and anyone can download it. It's huge, but uses only a small slice of itself for each word, so it costs far less to run than its size suggests.

1.05T params · 52B active · 1M context · open weights

docs.mistral.ai Scores & price Oct 9 issue

Sources

Mistral Large 4 on OpenRouter
Epoch AI benchmarks hub