MiMo-V2.6-Pro (Thinking): benchmark results
Provider: Xiaomi. Released 2026-09-21. Access: Open.
Aggregate Rating 1810 ± 24, rank #128 of 2257 rated models, from 19 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| VehicleArena - Basic Request Fulfillment | 90 | Score (0-100) | 100 |
| NYT Connections Older Models | 84 | Score (%) | 92.9 |
| VehicleArena - Multi-Agent Request Fulfillment | 88 | Score (0-100) | 87.5 |
| DuelLab GameBench 2 - Reasoning Variants | 70.77 | GameBench2 normalized score (0–100) | 82.4 |
| VehicleArena - Multi-Agent Arrival | 62.5 | Arrival rate (%) | 81.2 |
| Wolfram LLM Benchmarking Project | 58.1 | Correct Functionality (%) | 79.2 |
| Multi-turn Debate (Lechmazur) | 1600.3 | Bradley-Terry Rating | 75 |
| VehicleArena - Multi-Agent Cabin | 66.7 | Score (0-100) | 75 |
| VehicleArena - Basic Cabin | 57.5 | Score (0-100) | 62.5 |
| VehicleArena - Basic Arrival | 40 | Arrival rate (%) | 50 |
| VehicleArena - Basic Driving | 79.5 | Score (0-100) | 50 |
| VehicleArena - Multi-Agent Driving | 80 | Score (0-100) | 50 |
Interactive version: theaggregate.ai/model?slug=mimo-v2-6-pro-thinking · How It Works · Data refreshed daily, snapshot 2026-11-17.