ModelBeat
Models/Mistral Medium 3.5
All models
M

Mistral AI: Mistral Medium 3.5

mistralai/mistral-medium-3-5

Mistral Medium 3.5 is Mistral AI’s open-weights AI model, released Apr 28, 2026. It costs $1.50/M input and $7.50/M output tokens, and handles a 262K-token context window. Its strongest use case is agentic & tools, where it ranks in the 75th percentile of the models Model Beat tracks.

Mistral Medium 3.5 is a dense 128B instruction-following model from Mistral AI.

Context
262K
Input
$1.50 / 1M
Output
$7.50 / 1M
Released
Apr 28, 2026
Modalities
textimagefiletext
Strongest atAgentic & Tools top 25%

Benchmarks

Scores on standardized evaluations. Higher is better — and rank shows where Mistral Medium 3.5 lands among all models tracked on Model Beat.

37.0
Intelligence Index
Epoch AI
37th percentile of tracked models
19.0
Coding Index
Epoch AI
19th percentile of tracked models
75.0
Agentic Index
Epoch AI
75th percentile of tracked models

Reasoning

3 evals
GPQA Diamond74.8%

Graduate-level scientific reasoning

Humanity's Last Exam13.8%

Frontier of human expert knowledge

WeirdML43.7%

Novel ML problem-solving

Coding

2 evals
SciCode40.2%

Scientific research coding

WebDev Arena1265

Human-rated web development

Agentic & Tools

1 evals
τ²-bench94.2%

Tool-agent-user reliability

In the news

3

Changelog

1
  • Mistral Medium 3.5: Humanity's Last Exam improved from 12.8% to 13.8% (+8%)benchmark · Aug 6, 2026

Compare Mistral Medium 3.5 with

Frequently asked questions

What is Mistral Medium 3.5?

Mistral Medium 3.5 is a dense 128B instruction-following model from Mistral AI.

Who created Mistral Medium 3.5?

Mistral Medium 3.5 is an AI model developed by Mistral AI, part of the Mistral family, released Apr 28, 2026.

How much does Mistral Medium 3.5 cost?

Mistral Medium 3.5's list price from Mistral AI is $1.50 per million input tokens and $7.50 per million output tokens.

What is Mistral Medium 3.5's context window?

Mistral Medium 3.5 supports a context window of up to 262K tokens.

How does Mistral Medium 3.5 perform on benchmarks?

On standardized evaluations tracked by Epoch AI, Mistral Medium 3.5 scores 94.2% on τ²-bench, 40.2% on SciCode, 43.7% on WeirdML.

Is Mistral Medium 3.5 open source?

Mistral Medium 3.5 is released with open weights (Open weights (restricted use)), so the model can be downloaded and self-hosted.

Benchmark data from Epoch AI (CC BY) and Artificial Analysis; pricing, specs & descriptions from OpenRouter. Data updated Sep 10, 2026.