Meta · model
Muse Spark Open benchmark scores
As of 2026-10-29, Muse Spark Open (Meta) has measured scores on 23 benchmarks and estimated scores on 108 more.
| Benchmark | Score | Source |
|---|---|---|
| AA-GPQA Diamond | 89.0% | measured |
| AA-HLE | 43.9% | measured |
| AA-SciCode | 58.4% | measured |
| CritPt | 14.9% | measured |
| GDPval-AA | 45.1% | measured |
| GPQA Diamond | 89.4% | estimated ± 2.3 pp, medium confidence |
| HLE | 47.9% | estimated ± 9.7 pp, medium confidence |
| AA-LCR | 78.3% | measured |
| LiveCodeBench v6 | 98.9% | estimated ± 4.9 pp, medium confidence |
| SciCode | 58.0% | estimated ± 3.1 pp, high confidence |
| Terminal-Bench 2.1 | 79.1% | estimated ± 4.5 pp, high confidence |
| BrowseComp | 86.0% | estimated ± 3.9 pp, medium confidence |
| HLE w/ tools | 61.0% | measured |
| LongBench v2 | 61.8% | estimated ± 5.8 pp, low confidence |
| VITA-Bench | 51.7% | estimated ± 9.9 pp, low confidence |
| τ²-bench results | 98.5% | estimated ± 3.1 pp, medium confidence |
| AA Coding Index | 73.1% | estimated ± 2.5 pp, high confidence |
| AA-MMMU-Pro | 80.9% | measured |
| APEX-Agents | 41.2% | estimated ± 0.6 pp, low confidence |
| APEX-Agents-AA | 41.7% | estimated ± 3.6 pp, low confidence |
| BioMysteryBench (human-difficult) | 46.5% | estimated ± 6.8 pp, low confidence |
| DeepSearchQA | 89.3% | estimated ± 9.6 pp, medium confidence |
| SWE-bench Verified | 84.8% | estimated ± 4.6 pp, high confidence |
| BFCL v4 | 76.2% | estimated ± 2.7 pp, medium confidence |
| CyberGym | 83.5% | estimated ± 6.0 pp, low confidence |
| JobBench | 61.8% | measured |
| SWE-bench Pro | 72.4% | estimated ± 6.6 pp, medium confidence |
| WideResearch | 83.8% | estimated ± 1.2 pp, low confidence |
| AA Agentic Index | 39.0% | estimated ± 2.9 pp, high confidence |
| MMLU-Pro | 86.2% | estimated ± 4.1 pp, high confidence |
| Artificial Analysis Intelligence Index | 24.7% | measured |
| GPQA | 89.5% | estimated ± 6.1 pp, medium confidence |
| MMLU | 94.8% | estimated ± 0.7 pp, medium confidence |
| AA-Omniscience Accuracy | 42.0% | measured |
| MMMLU | 90.2% | estimated ± 0.7 pp, low confidence |
| MMMU | 87.0% | estimated ± 2.7 pp, medium confidence |
| AA MMLU-Pro | 87.5% | estimated ± 0.6 pp, high confidence |
| Gert Labs | 100.0% | estimated ± 7.3 pp, low confidence |
| AA-AnalystAgent | 39.1% | estimated ± 7.8 pp, low confidence |
| AA Harvey LAB v1.0 | 91.3% | estimated ± 1.7 pp, medium confidence |
| AA Tau3 Banking | 41.0% | estimated ± 4.0 pp, high confidence |
| AA Terminal-Bench 2.1 | 76.4% | measured |
| ApprenticeBench | 20.2% | estimated ± 5.6 pp, low confidence |
| ARC-AGI-1 | 84.9% | estimated ± 7.6 pp, medium confidence |
| ARC-AGI-2 | 71.1% | estimated ± 14.6 pp, low confidence |
| CursorBench 3.1 | 64.7% | estimated ± 3.4 pp, medium confidence |
| CursorBench 3.2 | 67.1% | estimated ± 5.3 pp, medium confidence |
| Terminal-Bench 3.0 | 13.4% | estimated ± 8.6 pp, low confidence |
| FrontierCode 1.1 Main | 38.9% | estimated ± 3.7 pp, low confidence |
| FrontierSWE v2 | 37.8% | estimated ± 14.6 pp, low confidence |
| MazeBench | 0.0% | measured |
| OfficeQA Pro | 63.4% | estimated ± 6.8 pp, medium confidence |
| OSWorld-Verified | 81.0% | estimated ± 8.1 pp, medium confidence |
| ParseBench | 67.2% | estimated ± 5.5 pp, medium confidence |
| Vals GPQA Diamond | 85.5% | estimated ± 5.3 pp, low confidence |
| Vals MMLU-Pro | 84.8% | estimated ± 2.1 pp, high confidence |
| Vals SWE-bench | 89.6% | estimated ± 7.5 pp, low confidence |
| Terminal-Bench 2.1 (Vals AI) | 63.6% | estimated ± 6.1 pp, medium confidence |
| AA AutomationBench | 49.7% | measured |
| CWE-Bench-AA | 59.7% | estimated ± 5.2 pp, medium confidence |
| AA Cyber Index | 39.7% | estimated ± 12.8 pp, low confidence |
| DeepsecBench-AA | 17.8% | estimated ± 5.2 pp, low confidence |
| GDP.pdf | 18.6% | measured |
| AA ITBench | 49.7% | estimated ± 4.4 pp, medium confidence |
| AA Terminal-Bench 4.0 | 8.1% | measured |
| Bug Hunt Bench | 22.2% | estimated ± 9.4 pp, low confidence |
| CursorBench 4.0 | 45.6% | estimated ± 3.5 pp, high confidence |
| CWE-bench v1 | 58.0% | estimated ± 2.7 pp, medium confidence |
| DeepSWE | 60.0% | measured |
| HLE w/o tools | 35.1% | estimated ± 4.6 pp, high confidence |
| OSWorld 2.0 | 19.4% | estimated ± 5.5 pp, low confidence |
| OSWorld 2.1 | 78.3% | estimated ± 2.8 pp, medium confidence |
| PostTrainBench v1.1 | 28.7% | estimated ± 4.5 pp, high confidence |
| Terminal-Bench 4.0 | 11.2% | estimated ± 5.7 pp, low confidence |
| Toolathlon-Verified | 72.8% | measured |
| OSWorld 2.1 Offline | 51.3% | estimated ± 7.9 pp, low confidence |
| ARC-AGI-3 | 1.4% | estimated ± 7.7 pp, medium confidence |
| ARC-AGI-3 Provider Adapter | 3.9% | estimated ± 7.5 pp, medium confidence |
| Hermes Bench | 68.2% | estimated ± 2.9 pp, high confidence |
| Hermes Index | 37.2% | estimated ± 3.1 pp, high confidence |
| CharXiv | 90.2% | estimated ± 3.6 pp, high confidence |
| CharXiv w/o tools | 83.6% | estimated ± 2.6 pp, high confidence |
| ExploitGym | 11.6% | estimated ± 3.3 pp, high confidence |
| AI2D_TEST | 94.1% | estimated ± 2.0 pp, low confidence |
| C-Eval | 93.0% | estimated ± 1.0 pp, medium confidence |
| Claw-Eval | 66.2% | estimated ± 5.6 pp, low confidence |
| CountBench | 98.7% | estimated ± 2.9 pp, low confidence |
| DeepPlanning | 54.3% | estimated ± 13.1 pp, low confidence |
| DynaMath | 85.5% | estimated ± 3.0 pp, low confidence |
| ERQA | 68.9% | estimated ± 2.9 pp, high confidence |
| MathVision | 93.6% | estimated ± 1.0 pp, medium confidence |
| MCP Atlas | 89.2% | measured |
| MLVU (M-Avg) | 87.2% | estimated ± 5.0 pp, high confidence |
| MMLU-Redux | 96.2% | estimated ± 0.8 pp, low confidence |
| MMMU-Pro | 81.4% | estimated ± 1.6 pp, high confidence |
| NL2Repo | 56.1% | estimated ± 6.2 pp, medium confidence |
| OmniDocBench 1.5 | 90.2% | estimated ± 2.0 pp, medium confidence |
| QwenClawBench | 74.1% | estimated ± 2.6 pp, medium confidence |
| RealWorldQA | 86.6% | estimated ± 2.7 pp, medium confidence |
| ScreenSpot Pro | 77.6% | estimated ± 11.2 pp, low confidence |
| SimpleVQA | 73.6% | measured |
| SuperGPQA | 84.9% | estimated ± 9.3 pp, medium confidence |
| Toolathlon | 55.9% | estimated ± 6.3 pp, medium confidence |
| V* | 100.0% | estimated ± 7.5 pp, low confidence |
| Video-MME (w/o subtitle) | 87.4% | estimated ± 1.7 pp, low confidence |
| Video-MME (with subtitle) | 88.8% | estimated ± 3.7 pp, low confidence |
| VideoMMMU | 86.9% | estimated ± 10.3 pp, low confidence |
| We-Math | 84.9% | estimated ± 6.7 pp, low confidence |
| HealthBench Hard | 31.2% | estimated ± 8.6 pp, low confidence |
| MedXpertQA (MM) | 77.2% | estimated ± 6.5 pp, low confidence |
| React Native Evals | 80.2% | estimated ± 4.7 pp, medium confidence |
| ResearchClawBench | 20.0% | estimated ± 1.8 pp, low confidence |
| Finance Agent v2 | 56.1% | estimated ± 3.8 pp, medium confidence |
| BioMysteryBench (human-solvable) | 87.5% | estimated ± 0.6 pp, medium confidence |
| DRACO | 80.1% | estimated ± 3.1 pp, medium confidence |
| HealthBench (raw) | 52.2% | estimated ± 6.0 pp, medium confidence |
| HealthBench (length-adjusted) | 53.9% | estimated ± 3.3 pp, medium confidence |
| HealthBench Professional | 58.3% | estimated ± 4.5 pp, medium confidence |
| HealthBench Professional (raw) | 62.6% | estimated ± 4.5 pp, high confidence |
| HLE-Verified | 15.7% | estimated ± 2.0 pp, low confidence |
| LABBench2 | 75.5% | estimated ± 1.8 pp, low confidence |
| STEMBench | 2.4% | estimated ± 0.2 pp, high confidence |
| SimpleQA | 48.3% | estimated ± 4.3 pp, low confidence |
| Agents' Last Exam | 37.0% | estimated ± 10.1 pp, low confidence |
| OpenHarmony Bench | 58.9% | estimated ± 2.4 pp, high confidence |
| BabyVision | 80.2% | measured |
| ZeroBench | 50.0% | measured |
| MRCRv2 | 85.5% | estimated ± 11.3 pp, low confidence |
| MMLongBench-Doc | 61.3% | estimated ± 2.6 pp, low confidence |
| MMMU-Pro w/ Python | 83.5% | estimated ± 2.0 pp, high confidence |
| SWE-Atlas Codebase QnA | 58.0% | measured |