MiniMax M1 80k: benchmark results
MiniMax M1 open hybrid-attention reasoning model at the 80K thinking budget. Provider: MiniMax. Released 2025-06-16. Access: Open.
Aggregate Rating 1603 ± 18, rank #432 of 1656 rated models, from 138 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| OpenCompass Math - College - Chinese (CompassBench 2507) | 56 | Score (%) | 91.7 |
| BenCzechMark - Factual Knowledge | 71.8 | Duel Win Score (%) | 90.9 |
| BenCzechMark - Reading Comprehension | 75.32 | Duel Win Score (%) | 90.9 |
| BenCzechMark | 70.71 | Average Duel Win Score (%) | 84.4 |
| BenCzechMark - Czech Math Reasoning | 70.45 | Duel Win Score (%) | 84.4 |
| OpenCompass LLM - Math (CompassBench 2507) | 59.6 | Score (%) | 83.3 |
| OpenCompass LLM - Math - English (CompassBench 2507) | 64.3 | Score (%) | 83.3 |
| OpenCompass Math - Competition - English (CompassBench 2507) | 60 | Score (%) | 83.3 |
| BenCzechMark - Sentiment | 72.73 | Duel Win Score (%) | 81.8 |
| AA LiveCodeBench | 71.11 | Pass@1 (%) | 81.4 |
| BenCzechMark - Czech Language Understanding | 66.67 | Duel Win Score (%) | 77.9 |
| OpenCompass Agent - Tool Use (CompassBench 2507) | 64.7 | Score (%) | 77.8 |
Interactive version: theaggregate.ai/model?slug=minimax-m1-80k · How It Works · Data refreshed daily, snapshot 2026-10-20.