Xiaomi · model
MiMo-V2.6-Flash benchmark scores
As of 2026-10-07, MiMo-V2.6-Flash (Xiaomi) has measured scores on 21 benchmarks and estimated scores on 82 more.
| Benchmark | Score | Source |
|---|---|---|
| AA-GPQA Diamond | 90.7% | estimated ± 6.8 pp, medium confidence |
| AA-HLE | 35.1% | measured |
| AA-SciCode | 51.3% | measured |
| Artificial Analysis Intelligence Index | 37.9% | measured |
| CritPt | 12.0% | measured |
| GDPval-AA | 55.5% | measured |
| GPQA Diamond | 89.3% | estimated ± 2.2 pp, high confidence |
| HLE | 45.8% | estimated ± 9.5 pp, medium confidence |
| AA-LCR | 74.3% | measured |
| LiveCodeBench v6 | 92.5% | estimated ± 3.3 pp, high confidence |
| AA-Omniscience Accuracy | 27.0% | measured |
| SciCode | 51.0% | estimated ± 3.2 pp, high confidence |
| Terminal-Bench 2.1 | 87.6% | measured |
| BrowseComp | 82.4% | estimated ± 1.8 pp, medium confidence |
| HLE w/ tools | 72.7% | estimated ± 5.0 pp, low confidence |
| LongBench v2 | 61.5% | estimated ± 5.9 pp, medium confidence |
| AA Coding Index | 75.5% | estimated ± 2.5 pp, high confidence |
| AA-MMMU-Pro | 73.1% | measured |
| APEX-Agents | 44.6% | estimated ± 10.0 pp, low confidence |
| APEX-Agents-AA | 41.5% | estimated ± 3.6 pp, low confidence |
| BioMysteryBench (human-difficult) | 32.8% | estimated ± 5.9 pp, low confidence |
| DeepSearchQA | 90.4% | estimated ± 8.4 pp, medium confidence |
| SWE-bench Verified | 89.2% | estimated ± 5.2 pp, low confidence |
| AutomationBench | 52.3% | measured |
| BFCL v4 | 77.9% | estimated ± 2.6 pp, medium confidence |
| CyberGym | 95.1% | measured |
| JobBench | 61.2% | measured |
| SWE-bench Pro | 61.2% | estimated ± 6.8 pp, medium confidence |
| WideResearch | 83.7% | estimated ± 1.2 pp, medium confidence |
| AA Agentic Index | 52.7% | estimated ± 1.7 pp, high confidence |
| MMLU-Pro | 85.6% | estimated ± 3.5 pp, high confidence |
| GPQA | 91.7% | estimated ± 3.2 pp, high confidence |
| Gert Labs | 73.4% | estimated ± 6.0 pp, low confidence |
| AA-AnalystAgent | 42.5% | estimated ± 6.7 pp, low confidence |
| AA EnterpriseOps-Gym | 31.9% | estimated ± 5.1 pp, low confidence |
| ApprenticeBench | 32.7% | estimated ± 4.0 pp, low confidence |
| ARC-AGI-1 | 92.4% | estimated ± 5.4 pp, medium confidence |
| ARC-AGI-2 | 70.3% | estimated ± 14.2 pp, low confidence |
| CursorBench 3.1 | 47.2% | estimated ± 3.4 pp, low confidence |
| CursorBench 3.2 | 27.1% | estimated ± 3.6 pp, low confidence |
| Terminal-Bench 3.0 | 22.2% | estimated ± 9.1 pp, low confidence |
| FrontierCode 1.1 Main | 46.6% | estimated ± 3.7 pp, low confidence |
| FrontierSWE v2 | 15.6% | estimated ± 14.6 pp, low confidence |
| OfficeQA Pro | 50.3% | estimated ± 7.2 pp, low confidence |
| OSWorld-Verified | 80.8% | measured |
| Vals GPQA Diamond | 91.1% | estimated ± 7.5 pp, medium confidence |
| Vals LiveCodeBench | 77.4% | estimated ± 6.1 pp, low confidence |
| Vals MMLU-Pro | 84.4% | estimated ± 2.2 pp, high confidence |
| Vals SWE-bench | 80.9% | estimated ± 7.6 pp, low confidence |
| Terminal-Bench 2.1 (Vals AI) | 76.3% | estimated ± 4.6 pp, medium confidence |
| AA AutomationBench | 56.7% | estimated ± 5.0 pp, medium confidence |
| GDP.pdf | 13.9% | estimated ± 3.4 pp, high confidence |
| AA ITBench | 46.3% | estimated ± 3.9 pp, low confidence |
| AA Tau3 Banking | 47.5% | estimated ± 5.8 pp, medium confidence |
| AA Terminal-Bench 2.1 | 84.9% | estimated ± 3.5 pp, high confidence |
| AA Terminal-Bench 4.0 | 26.2% | estimated ± 4.7 pp, high confidence |
| CursorBench 4.0 | 31.2% | estimated ± 3.5 pp, medium confidence |
| CWE-bench v1 | 59.4% | estimated ± 6.9 pp, medium confidence |
| DeepSWE | 67.9% | measured |
| HLE w/o tools | 32.7% | estimated ± 5.7 pp, medium confidence |
| OSWorld 2.0 | 15.7% | estimated ± 3.6 pp, high confidence |
| PostTrainBench v1.1 | 36.7% | estimated ± 4.8 pp, high confidence |
| ProgramBench | 26.0% | measured |
| Terminal-Bench 4.0 | 28.8% | measured |
| Toolathlon-Verified | 73.6% | measured |
| CharXiv | 76.9% | estimated ± 7.5 pp, low confidence |
| CharXiv w/o tools | 78.4% | estimated ± 2.4 pp, medium confidence |
| ExploitBench | 25.3% | measured |
| ExploitGym | 6.0% | measured |
| C-Eval | 95.4% | estimated ± 1.0 pp, low confidence |
| Claw-Eval | 66.2% | estimated ± 5.6 pp, low confidence |
| MathVision | 82.1% | estimated ± 4.9 pp, high confidence |
| MCP Atlas | 81.0% | estimated ± 7.8 pp, medium confidence |
| MMLU-Redux | 95.0% | estimated ± 1.1 pp, medium confidence |
| MMMU-Pro | 76.1% | estimated ± 4.4 pp, medium confidence |
| NL2Repo | 51.8% | estimated ± 6.5 pp, medium confidence |
| QwenClawBench | 100.0% | estimated ± 6.0 pp, low confidence |
| ScreenSpot Pro | 71.8% | estimated ± 10.7 pp, low confidence |
| SuperGPQA | 72.0% | estimated ± 7.5 pp, medium confidence |
| Toolathlon | 55.9% | estimated ± 6.3 pp, medium confidence |
| ERQA | 55.4% | estimated ± 3.9 pp, high confidence |
| HealthBench Hard | 32.8% | estimated ± 7.7 pp, low confidence |
| MedXpertQA (MM) | 66.8% | estimated ± 6.5 pp, low confidence |
| MMLU-Pro (Arcee) | 100.0% | estimated ± 4.0 pp, low confidence |
| React Native Evals | 77.2% | estimated ± 4.7 pp, medium confidence |
| ResearchClawBench | 19.9% | estimated ± 2.2 pp, low confidence |
| ARC-AGI-3 | 1.7% | estimated ± 3.4 pp, medium confidence |
| BioMysteryBench (human-solvable) | 85.1% | estimated ± 0.6 pp, low confidence |
| DRACO | 84.3% | estimated ± 3.0 pp, medium confidence |
| HealthBench (raw) | 44.4% | estimated ± 6.4 pp, low confidence |
| HealthBench (length-adjusted) | 53.9% | estimated ± 3.7 pp, low confidence |
| HealthBench Professional | 57.7% | estimated ± 4.8 pp, low confidence |
| HealthBench Professional (raw) | 55.1% | estimated ± 5.0 pp, low confidence |
| HLE-Verified | 21.4% | estimated ± 6.0 pp, low confidence |
| LABBench2 | 77.7% | estimated ± 2.4 pp, low confidence |
| MMMU | 81.5% | estimated ± 0.9 pp, high confidence |
| Agents' Last Exam | 27.6% | measured |
| OpenHarmony Bench | 58.0% | estimated ± 2.4 pp, high confidence |
| SEC-Bench Pro | 47.5% | measured |
| MMLU | 89.8% | estimated ± 1.1 pp, medium confidence |
| MMMU-Pro w/ Python | 77.4% | estimated ± 1.6 pp, high confidence |
| RealWorldQA | 83.9% | estimated ± 1.2 pp, low confidence |
| OmniDocBench 1.5 | 28.5% | estimated ± 6.4 pp, low confidence |