Alibaba (Qwen) · model
Qwen3.8-27B benchmark scores
As of 2026-10-07, Qwen3.8-27B (Alibaba (Qwen)) has measured scores on 44 benchmarks and estimated scores on 79 more.
| Benchmark | Score | Source |
|---|---|---|
| AA-GPQA Diamond | 90.5% | measured |
| AA-HLE | 33.9% | measured |
| AA-SciCode | 46.6% | measured |
| Artificial Analysis Intelligence Index | 33.7% | measured |
| CritPt | 5.4% | measured |
| GDPval-AA | 46.2% | measured |
| GPQA Diamond | 89.2% | measured |
| HLE | 30.8% | measured |
| IFBench | 79.5% | measured |
| AA-LCR | 82.0% | measured |
| LiveCodeBench v6 | 90.3% | measured |
| AA-Omniscience Accuracy | 15.6% | measured |
| SciCode | 44.7% | estimated ± 2.9 pp, medium confidence |
| Terminal-Bench 2.1 | 73.0% | measured |
| BrowseComp | 83.8% | estimated ± 3.1 pp, high confidence |
| HLE w/ tools | 55.3% | estimated ± 6.1 pp, medium confidence |
| IFEval | 92.8% | estimated ± 3.5 pp, high confidence |
| LongBench v2 | 61.5% | estimated ± 5.9 pp, low confidence |
| τ²-bench results | 94.3% | estimated ± 2.4 pp, medium confidence |
| AA Coding Index | 68.1% | measured |
| AA-MMMU-Pro | 76.3% | measured |
| APEX-Agents | 38.6% | estimated ± 10.0 pp, low confidence |
| APEX-Agents-AA | 31.5% | estimated ± 3.6 pp, medium confidence |
| BioMysteryBench (human-difficult) | 22.7% | estimated ± 5.9 pp, low confidence |
| DeepSearchQA | 92.9% | estimated ± 4.5 pp, high confidence |
| SWE-bench Verified | 78.3% | estimated ± 2.4 pp, high confidence |
| AutomationBench | 38.9% | estimated ± 7.0 pp, low confidence |
| BFCL v4 | 77.1% | estimated ± 2.6 pp, medium confidence |
| CyberGym | 81.3% | estimated ± 5.3 pp, medium confidence |
| JobBench | 33.4% | measured |
| SWE-bench Pro | 61.7% | measured |
| WideResearch | 77.6% | estimated ± 1.2 pp, medium confidence |
| AA Agentic Index | 46.5% | measured |
| MMLU-Pro | 85.2% | estimated ± 2.5 pp, medium confidence |
| GPQA | 89.2% | measured |
| AA-IFBench | 80.3% | estimated ± 8.5 pp, medium confidence |
| Gert Labs | 78.5% | estimated ± 7.2 pp, low confidence |
| AA-AnalystAgent | 34.2% | estimated ± 3.6 pp, medium confidence |
| AA EnterpriseOps-Gym | 44.2% | measured |
| AA Harvey LAB | 93.2% | estimated ± 1.2 pp, medium confidence |
| ApprenticeBench | 29.5% | estimated ± 5.2 pp, low confidence |
| ARC-AGI-1 | 87.9% | estimated ± 2.7 pp, high confidence |
| ARC-AGI-2 | 52.6% | estimated ± 10.1 pp, low confidence |
| CursorBench 3.1 | 35.6% | estimated ± 3.4 pp, low confidence |
| CursorBench 3.2 | 64.9% | estimated ± 3.3 pp, medium confidence |
| Terminal-Bench 3.0 | 4.9% | estimated ± 6.9 pp, low confidence |
| FrontierCode 1.1 Main | 24.3% | estimated ± 3.7 pp, low confidence |
| FrontierSWE v2 | 18.2% | estimated ± 9.0 pp, medium confidence |
| OfficeQA Pro | 50.3% | estimated ± 7.2 pp, low confidence |
| OSWorld-Verified | 84.3% | measured |
| Vals GPQA Diamond | 88.9% | measured |
| Vals LiveCodeBench | 84.0% | measured |
| Vals MMLU-Pro | 84.3% | measured |
| Vals SWE-bench | 86.0% | measured |
| Terminal-Bench 2.1 (Vals AI) | 58.4% | measured |
| VulcanBench v3 | 82.6% | measured |
| AA AutomationBench | 48.2% | measured |
| GDP.pdf | 16.6% | measured |
| AA ITBench | 50.0% | estimated ± 5.7 pp, low confidence |
| AA Tau3 Banking | 48.0% | measured |
| AA Terminal-Bench 2.1 | 79.8% | measured |
| AA Terminal-Bench 4.0 | 5.6% | measured |
| CursorBench 4.0 | 30.4% | estimated ± 1.6 pp, medium confidence |
| CWE-bench v1 | 46.6% | estimated ± 6.0 pp, medium confidence |
| DeepSWE | 42.2% | measured |
| HLE w/o tools | 30.8% | measured |
| OSWorld 2.0 | 19.6% | estimated ± 3.6 pp, high confidence |
| PostTrainBench v1.1 | 29.1% | estimated ± 2.7 pp, high confidence |
| Terminal-Bench 4.0 | 9.5% | estimated ± 5.1 pp, low confidence |
| Terminal-Bench-Science 0.1 | 0.0% | estimated ± 4.1 pp, low confidence |
| Toolathlon-Verified | 73.2% | estimated ± 1.3 pp, low confidence |
| Vibe Code Bench | 48.0% | estimated ± 10.3 pp, low confidence |
| CharXiv | 90.2% | measured |
| CharXiv w/o tools | 83.7% | measured |
| ExploitGym | 14.3% | estimated ± 5.0 pp, low confidence |
| C-Eval | 92.9% | estimated ± 0.8 pp, medium confidence |
| Claw-Eval | 65.0% | estimated ± 5.6 pp, low confidence |
| MathVision | 90.0% | measured |
| MCP Atlas | 82.9% | estimated ± 2.7 pp, medium confidence |
| MMLU-Redux | 91.4% | estimated ± 1.1 pp, medium confidence |
| MMMU-Pro | 80.3% | estimated ± 1.4 pp, medium confidence |
| NL2Repo | 42.3% | measured |
| QwenClawBench | 100.0% | estimated ± 6.0 pp, low confidence |
| ScreenSpot Pro | 74.5% | estimated ± 4.5 pp, medium confidence |
| SuperGPQA | 57.4% | estimated ± 7.5 pp, medium confidence |
| Terminal-Bench 2.0 | 65.3% | estimated ± 5.9 pp, medium confidence |
| Toolathlon | 55.9% | estimated ± 6.3 pp, medium confidence |
| V* | 96.7% | estimated ± 5.1 pp, low confidence |
| VideoMMMU | 87.2% | estimated ± 1.5 pp, high confidence |
| ERQA | 65.5% | measured |
| HealthBench Hard | 0.4% | estimated ± 5.1 pp, low confidence |
| LiveCodeBench Pro | 85.0% | estimated ± 6.0 pp, low confidence |
| MedXpertQA (MM) | 76.3% | estimated ± 4.2 pp, medium confidence |
| MedXpertQA (Text) | 51.5% | estimated ± 4.5 pp, medium confidence |
| MMLU-Pro (Arcee) | 86.6% | estimated ± 3.4 pp, medium confidence |
| React Native Evals | 92.0% | estimated ± 2.4 pp, low confidence |
| ResearchClawBench | 20.7% | estimated ± 2.0 pp, medium confidence |
| SWE-Rebench | 58.1% | estimated ± 8.9 pp, low confidence |
| ARC-AGI-3 | 0.0% | estimated ± 2.1 pp, medium confidence |
| BioMysteryBench (human-solvable) | 81.8% | estimated ± 0.6 pp, low confidence |
| DRACO | 80.5% | estimated ± 3.0 pp, medium confidence |
| FrontierCode 1.1 Extended | 41.2% | estimated ± 2.4 pp, low confidence |
| HealthBench (raw) | 43.2% | estimated ± 6.4 pp, low confidence |
| HealthBench (length-adjusted) | 53.9% | estimated ± 3.7 pp, low confidence |
| HealthBench Professional | 51.2% | estimated ± 1.3 pp, low confidence |
| HealthBench Professional (raw) | 54.0% | estimated ± 5.0 pp, low confidence |
| HLE-Verified | 21.9% | estimated ± 2.0 pp, low confidence |
| LABBench2 | 79.9% | estimated ± 1.8 pp, low confidence |
| MMMU | 84.1% | estimated ± 0.9 pp, high confidence |
| LiveCodeBench | 100.0% | estimated ± 8.9 pp, low confidence |
| Agents' Last Exam | 42.9% | measured |
| OpenHarmony Bench | 51.0% | estimated ± 2.4 pp, high confidence |
| BabyVision | 65.7% | measured |
| MMVU | 77.7% | estimated ± 2.9 pp, medium confidence |
| SimpleVQA | 67.6% | estimated ± 7.4 pp, low confidence |
| ZeroBench | 16.9% | estimated ± 4.7 pp, low confidence |
| MMMLU | 89.5% | estimated ± 1.1 pp, medium confidence |
| MMLU | 82.6% | estimated ± 1.1 pp, medium confidence |
| MMMU-Pro w/ Python | 79.5% | estimated ± 1.6 pp, high confidence |
| MathVision w/ Python | 94.6% | measured |
| RealWorldQA | 85.9% | measured |
| Video-MME (with subtitle) | 89.5% | estimated ± 0.9 pp, medium confidence |
| OmniDocBench 1.5 | 91.1% | measured |