Mistral AI·released 2024-02-26
Scored on 4 benchmarks. 75% of its scores are independently reproduced — the rest are self-reported or unverified.
Every score, ranked, colored by evidence status.
| Benchmark | Score | Evidence status | Measured |
|---|---|---|---|
| MMLU | 58.4accuracy (%) | unverified· optimized | 2024-02-26 |
| MATH level 5 | 24.46accuracy (%) | reproduced· optimized | 2024-02-26 |
| GPQA diamond | 18.35accuracy (%) | reproduced· optimized | 2024-02-26 |
| OTIS Mock AIME 2024-2025 | 1.85accuracy (%) | reproduced· optimized | 2024-02-26 |