Magistral Medium 1

Mistral

Verified against Artificial Analysis · Jul 13, 2026

Capability

percentile index

Price /M

$0 / $0

in / out

Context

—

Avg score

38.6

11 benchmarks

AA Intelligence

12.5

index

Weights

Closed

Released

Jun 10, 2025

About

Magistral Medium 1 is a proprietary model in the Magistral family from Mistral. Benchmarked on 11 evals, averaging 38.6.

Benchmark scores · 11

avg 38.6 — higher is better

MATHartificial-analysis · 2026-07

91.7

MMLU-Proartificial-analysis · 2026-07

75.3

GPQA Diamondartificial-analysis · 2026-07

67.9

LiveCodeBenchartificial-analysis · 2026-07

52.7

AIMEartificial-analysis · 2026-07

40.3

SciCodeartificial-analysis · 2026-07

29.7

IFBenchartificial-analysis · 2026-07

25.1

τ²-benchartificial-analysis · 2026-07

23.1

Humanity's Last Examartificial-analysis · 2026-07

9.5

Terminal-Bench Hardartificial-analysis · 2026-07

9.1

AA Long-Context Reasoningartificial-analysis · 2026-07

Details

PublisherMistral

FamilyMagistral

Parameters—

License—

WeightsClosed

ReleasedJun 10, 2025

History

full ledger →

Capability over time

Append-only ledger — every observed change to this model's numbers.

API providers

Artificial Analysis$0/M