Magistral Small: benchmark results
Mistral's Apache-2.0 open 24B reasoning model, SFT-distilled from Magistral Medium traces plus RL (June 2025). Provider: Mistral. Released 2025-06-10. Access: Open.
Aggregate Rating 1447 ± 13, rank #1016 of 1656 rated models, from 165 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| Greek MMLU - Greek-Specific | 78.83 | Accuracy (%) | 81.5 |
| Greek MMLU | 71.27 | Accuracy (%) | 80.2 |
| Greek MMLU - Humanities | 73.15 | Accuracy (%) | 79 |
| BRIDGE Medical Leaderboard - Few-Shot | 49.15 | Average Performance (%) | 78.4 |
| Medmarks - LongHealth Task 2 | 88.96 | Score (%) | 78.4 |
| Greek MMLU - STEM | 69.06 | Accuracy (%) | 77.8 |
| Greek MMLU - Social Sciences | 74.61 | Accuracy (%) | 77.8 |
| SpeechMap Compliance | 78.1 | % Requests Completed | 75.5 |
| Greek MMLU - Other | 65.94 | Accuracy (%) | 72.8 |
| BRIDGE Medical Leaderboard - Zero-Shot | 37.56 | Average Performance (%) | 60.4 |
| AA LiveCodeBench | 51.43 | Pass@1 (%) | 57.7 |
| AA-Omniscience Hallucination Rate | 76.13 | Hallucination Rate (%) | 55.2 |
Interactive version: theaggregate.ai/model?slug=magistral-small · How It Works · Data refreshed daily, snapshot 2026-10-20.