Benchmark Almanac

Magistral Medium 1.0

Magistral Medium 1.0.

Mistral AI Released June 10th, 2025 Family Magistral The lab's model card Compare with others
25
Benchmarks with a score
0
Ranked first, out of five or more models
0
In the top three
Mistral AI Jun 2025

Magistral Medium 1.0

Mid-size

134.0 #237 of 404
Self-Reported Index, from the lab's own numbers. No Epoch Capabilities Index score yet.
  • MMMU-Pro 57.9% #69 /90
  • LiveCodeBench v5 59.4% #27 /47
  • Molecular Biology Capabilities Test 30.8% #85 /90

Where it ranks

Its best score on each benchmark, and where that ranks among every model with a score there. We rank each model once, at its best setting.

Knowledge and language

BenchmarkScoreRankSource
HUMAINE 31 #13 of 50 Official leaderboard
IFEval (Mistral internal version) 87.4% #1 of 2 Lab model card
LMArena (Elo) 1,305 #227 of 335 Official leaderboard
Phare 62.3% #43 of 62 Official leaderboard
Humanity's Last Exam 9.00% #143 of 154 Lab model card

Math and science

BenchmarkScoreRankSource
AIME 2024 90% #18 of 88 Lab model card
Misguided Attention 51.7% #17 of 45 Official leaderboard
AIME 2025 83.3% #79 of 143 Lab model card
GPQA Diamond 70.8% #180 of 302 Lab model card
MATH-500 94.3% #44 of 70 Lab model card
ForecastBench 0.16 #70 of 82 Official leaderboard
ARC-AGI 6.12% #87 of 95 Official leaderboard
ARC-AGI-2 0% #97 of 98 Official leaderboard

Coding

BenchmarkScoreRankSource
LiveCodeBench v5 59.4% #27 of 47 Lab model card
Aider Polyglot 47.1% #23 of 30 Lab model card
LiveCodeBench v6 50.3% #81 of 90 Lab model card

Agents and tools

BenchmarkScoreRankSource
SnakeBench 26 #72 of 188 Official leaderboard
Galileo Agent Leaderboard v2 32% #16 of 21 Official leaderboard
Mistral internal function calling benchmark 87.4% #1 of 1 Lab model card

Professional work

BenchmarkScoreRankSource
Vals Finance Agent v1 7.46% #42 of 47 Official leaderboard

Vision, audio and robotics

BenchmarkScoreRankSource
MMMU 70% #64 of 103 Lab model card
MathVista 70.1% #47 of 70 Lab model card
MMMU-Pro 57.9% #69 of 90 Lab model card

Safety and security

BenchmarkScoreRankSource
Molecular Biology Capabilities Test 30.8% #85 of 90 Official leaderboard
Human Pathogen Capabilities Test 22.9% #84 of 86 Official leaderboard

Its lab's numbers put it at 134.0, number 237 of 404

The Self-Reported Index fits one score to the 11 capability results Mistral AI reported for this model, on the scale of the Epoch Capabilities Index. Labs choose which benchmarks to show and how to run them, so read it as the lab's claim, not an independent measurement. The 90% interval runs from 131.7 to 136.3. How we compute it.

Its launch post

The benchmarks the lab chose to headline, with the scores it gave.

Its system card

Every result in the longer documents the lab published with it. 9 of them aren't in the launch post.