Mistral Medium 3.1

Released

Mistral Medium 3.1 is an AI model released by Mistral on Tuesday, Aug 12 2025, 13 days after Codestral 25.08. It has a 128k token context windowHow much information the model can work with in one request, measured in tokens — pieces of words or other data. This includes your question, attachments, conversation and reply. Limits can vary by model variant and provider.. Benchmark results (shown below) cover GPQA Diamond, IFBench, MMLU-Pro, and MMMU-Pro.

Benchmarks

Reasoning & science

GPQA DiamondScience — Graduate-level science questions in biology, physics, and chemistry — hard enough that subject-matter PhDs score around 65%. Higher is better.
65.7%
#95 of 125
IFBenchInstruction following — Tests whether the AI can follow detailed instructions and satisfy multiple constraints at once. Higher is better.
40.8%
#7 of 8

Knowledge work

MMLU-ProKnowledge — accuracy on the ten-option MMLU-Pro benchmark
76.8%
#14 of 28

Multimodal

MMMU-ProMultimodal reasoning — A tougher version of MMMU — college-level questions that mix images, diagrams, and text together. Higher is better.
43.8%
#33 of 33

Compare Mistral Medium 3.1 with

Mistral Medium 3.1

About

Mistral Medium 3.1 was released on August 12, 2025 as an update to the commercial Medium 3 model. Mistral described improvements in performance and tone while continuing the line's focus on professional text, coding and multimodal applications.

The model supported a 128K-token context through hosted and enterprise offerings. It was a distinct checkpoint from the May Medium 3 release, even though both occupied the same product tier. The benchmark figures recorded here come from Mistral's later comparison published with Small 4 and should be understood as that evaluation of Medium 3.1, rather than measurements inherited from its predecessor.

Frequently asked questions

Mistral Medium 3.1 was released by Mistral on Tuesday, Aug 12 2025.

All Mistral releases

45 tracked