MiniMax-M2.7: benchmark results
MiniMax's open self-evolving M2.7 agentic MoE (230B total, 10B active). Provider: MiniMax. Released 2026-03-18. Access: Open.
Unified ELO 1627 ± 1, rank #188 of 1392 rated models, from 294 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| ClawSentry - Reviewer Replaceability | 27.78 | Task Success Rate (%) | 100 |
| LLM Stats (Multi-SWE-Bench) | 52.7 | Score (%) | 100 |
| MERA - LCS | 85.6 | Accuracy (%) | 99 |
| Enkrypt AI - Insecure Code Risk | 0.89 | Risk Score | 98.9 |
| Enkrypt AI - Risk Score | 13.08 | Risk Score | 98.5 |
| AA Omniscience - Software Engineering (SWE) - Swift | 64 | Accuracy (%) | 96.1 |
| Enkrypt AI - Bias Risk | 45.48 | Risk Score | 95.1 |
| AA IFBench | 75.71 | Accuracy (%) | 93.9 |
| MERA - MathLogicQA | 99.3 | Accuracy (%) | 93.8 |
| CritPt | 60 | Accuracy (self-reported) | 91.9 |
| MERA - ruMMLU | 84.53 | Accuracy (%) | 91.9 |
| AA Omniscience - Software Engineering (SWE) - Go | 38 | Accuracy (%) | 91.5 |
Interactive version: theaggregate.ai/model?slug=minimax-m2-7 · How It Works · Data refreshed daily, snapshot 2026-09-05.