Magistral Medium (Thinking) — benchmark results
Provider: Mistral. Released 2025-06-10. Access: API.
Unified ELO 1519 ± 47, rank #795 of 1839 rated models, from 9 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| CritPt | 0.3 | Accuracy (self-reported) | 65 |
| SpeechMap Compliance | 71.9 | % Requests Completed | 64.5 |
| SnakeBench | 20.7 | TrueSkill Rating | 44.8 |
| AI Chess Leaderboard (Reasoning) | 578 | Elo | 31.2 |
| LLM2014 Logic 2025-06 | 27.42 | Median Score | 22.2 |
| ARC-AGI-1 | 6.12 | Accuracy (%) | 8.6 |
| ARC-AGI-2 | 0 | Accuracy (%) | 4.8 |
| LLM2014 Logic 2025-08 | 14.43 | Median Score | 0 |
| LLM2014 Logic 2025-09 | 9.62 | Median Score | 0 |
Interactive version: theaggregate.ai/model?slug=magistral-medium-thinking · How It Works · Data refreshed daily, snapshot 2026-08-05.