MiMo-V2-Pro — benchmark results
Xiaomi's flagship MiMo-V2 Pro MoE model, served API-only. Provider: Xiaomi. Released 2026-03-18. Access: Open.
Unified ELO 1732 ± 19, rank #218 of 1776 rated models, from 83 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| Position Bias (Lechmazur) | 19.8 | Order Flip % (lower is better) | 100 |
| AA TAU-2 Bench | 95.03 | Accuracy (%) | 93.8 |
| BenchLM | 67.8 | Overall Score | 91.9 |
| AA Omniscience | 4.92 | Score | 91.1 |
| Artificial Analysis Intelligence Index | 40.29 | Intelligence Index | 90.8 |
| AA Humanity's Last Exam | 28.31 | Accuracy (%) | 88.7 |
| AA GPQA Diamond | 86.97 | Accuracy (%) | 88.1 |
| AA Terminal-Bench Hard | 40.91 | Accuracy (%) | 87.5 |
| BenchTable | 69.7 | Total Score (%) | 85.9 |
| Chatbot Arena (Text) | 1448 | Elo | 85.1 |
| AA Omniscience - Software Engineering (SWE) - Swift | 64 | Accuracy (%) | 84.9 |
| AA IFBench | 68.84 | Accuracy (%) | 82.8 |
Interactive version: theaggregate.ai/model?slug=mimo-v2-pro · How the rankings work · Data refreshed daily, snapshot 2026-07-22.