Open-weight reasoning model accepting text, image, audio, and video, with text output. Independent coverage is partial; API specifications are reported by Artificial Analysis.
SI v3 uses weighted percentages from one Artificial Analysis configuration. Ranking requires all five benchmarks. Partial scores average only available weights. Scores are not comparable to SI v2 or a probability of reaching singularity.
Compared with 58 models released within six months of it, across 3 benchmarks where scores spread out.
| Benchmark | Score | Rank |
|---|---|---|
HLEArtificial Analysis Challenging multidisciplinary questions evaluated by Artificial Analysis | 49.4% | #10 / 95 |
| Benchmark / source | Score |
|---|---|
HLE Artificial AnalysisMiMo-V2.6-Pro | 49.4% |
Terminal-Bench 4.0 Artificial AnalysisMiMo-V2.6-Pro | 34.8% |
SciCode Artificial AnalysisMiMo-V2.6-Pro | 60.9% |
AA-LCR Artificial AnalysisMiMo-V2.6-Pro | 86.3% |
CritPt Artificial AnalysisMiMo-V2.6-Pro | 26.6% |
DeepSWE 1.1 Provider reportXiaomi MiMo-V2.6 Pro release evaluation; harnesses as documented in model card | 71.9% |
ProgramBench Provider reportXiaomi MiMo-V2.6 Pro release evaluation; harnesses as documented in model card | 26.5% |
MiMo Code Bench Provider reportXiaomi MiMo-V2.6 Pro release evaluation; harnesses as documented in model card | 63.2% |
AutomationBench 1.0.6 Provider reportXiaomi MiMo-V2.6 Pro release evaluation; harnesses as documented in model card | 53.1% |
Toolathlon-Verified Provider reportXiaomi MiMo-V2.6 Pro release evaluation; harnesses as documented in model card | 76.9% |
GDPval-AA 2.1 Provider reportXiaomi MiMo-V2.6 Pro release evaluation; harnesses as documented in model card | 1673 Elo |
Agents' Last Exam Provider reportXiaomi MiMo-V2.6 Pro release evaluation; harnesses as documented in model card | 31.6% |
Terminal-Bench 4.0 Provider reportXiaomi MiMo-V2.6 Pro release evaluation; harnesses as documented in model card | 34.9% |
Terminal-Bench 2.1 Provider reportXiaomi MiMo-V2.6 Pro release evaluation; harnesses as documented in model card | 89.9% |
OSWorld-Verified Provider reportXiaomi MiMo-V2.6 Pro release evaluation; harnesses as documented in model card | 82% |
JobBench Provider reportXiaomi MiMo-V2.6 Pro release evaluation; harnesses as documented in model card | 62% |
CyberGym Provider reportXiaomi MiMo-V2.6 Pro release evaluation; harnesses as documented in model card | 94% |
MiMo Cyber Bench Provider reportXiaomi MiMo-V2.6 Pro release evaluation; harnesses as documented in model card | 80.2% |
ExploitGym Provider reportXiaomi MiMo-V2.6 Pro release evaluation; harnesses as documented in model card | 17.8% |
ExploitBench Provider reportXiaomi MiMo-V2.6 Pro release evaluation; harnesses as documented in model card | 47.9% |
SEC Bench Pro Provider reportXiaomi MiMo-V2.6 Pro release evaluation; harnesses as documented in model card | 66.3% |
MiMo VisualCoding Provider reportXiaomi MiMo-V2.6 Pro release evaluation; harnesses as documented in model card | 72.3% |