Other · model
Hy4 preview benchmark scores
As of 2026-11-21, Hy4 preview (Other) has measured scores on 32 benchmarks and estimated scores on 114 more.
| Benchmark | Score | Source |
|---|---|---|
| AA-GPQA Diamond | 90.9% | estimated ± 2.7 pp, high confidence |
| AA-HLE | 41.5% | estimated ± 3.9 pp, high confidence |
| AA-SciCode | 6.9% | estimated ± 3.0 pp, low confidence |
| Artificial Analysis Intelligence Index | 34.0% | estimated ± 5.6 pp, medium confidence |
| CritPt | 16.9% | measured |
| GDPval-AA | 41.0% | estimated ± 4.6 pp, medium confidence |
| GPQA Diamond | 92.3% | measured |
| HLE | 55.4% | measured |
| LiveCodeBench v6 | 94.6% | estimated ± 1.7 pp, medium confidence |
| AA-Omniscience Accuracy | 41.5% | estimated ± 7.7 pp, medium confidence |
| SciCode | 49.7% | estimated ± 5.9 pp, medium confidence |
| Terminal-Bench 2.1 | 85.4% | measured |
| BrowseComp | 90.8% | estimated ± 1.9 pp, medium confidence |
| HLE w/ tools | 55.4% | measured |
| VITA-Bench | 49.6% | estimated ± 9.9 pp, low confidence |
| τ²-bench results | 97.9% | estimated ± 3.1 pp, high confidence |
| AA Coding Index | 74.4% | estimated ± 2.5 pp, high confidence |
| AA-MMMU-Pro | 82.6% | estimated ± 2.3 pp, high confidence |
| APEX-Agents | 37.1% | measured |
| APEX-Agents-AA | 41.7% | estimated ± 3.6 pp, low confidence |
| BioMysteryBench (human-difficult) | 47.2% | estimated ± 6.8 pp, low confidence |
| DeepSearchQA | 91.0% | estimated ± 7.3 pp, low confidence |
| SWE-bench Verified | 81.8% | estimated ± 2.3 pp, high confidence |
| AutomationBench | 32.1% | measured |
| BFCL v4 | 77.0% | estimated ± 4.7 pp, low confidence |
| CyberGym | 78.4% | measured |
| JobBench | 61.7% | measured |
| SkillsBench | 62.9% | measured |
| SWE-bench Pro | 65.7% | measured |
| WideResearch | 83.9% | measured |
| AA Agentic Index | 50.6% | estimated ± 3.3 pp, medium confidence |
| MMLU-Pro | 85.7% | estimated ± 2.5 pp, low confidence |
| FailureSensorIQ | 66.4% | estimated ± 1.7 pp, high confidence |
| GPQA | 92.3% | measured |
| MMLU | 88.1% | estimated ± 4.2 pp, medium confidence |
| Vals GPQA Diamond | 91.4% | estimated ± 2.1 pp, high confidence |
| Vals LiveCodeBench | 86.5% | estimated ± 4.7 pp, medium confidence |
| Vals MMLU-Pro | 88.3% | estimated ± 0.9 pp, high confidence |
| MMMLU | 90.5% | estimated ± 0.7 pp, low confidence |
| Chess Puzzles | 42.6% | estimated ± 10.8 pp, low confidence |
| AA MMLU-Pro | 87.2% | estimated ± 1.2 pp, low confidence |
| Gert Labs | 67.4% | estimated ± 5.5 pp, medium confidence |
| AA-AnalystAgent | 47.1% | estimated ± 3.7 pp, medium confidence |
| AA EnterpriseOps-Gym | 46.7% | estimated ± 1.2 pp, medium confidence |
| AA Harvey LAB v1.0 | 89.6% | estimated ± 1.7 pp, medium confidence |
| AA Tau3 Banking | 39.4% | estimated ± 3.3 pp, medium confidence |
| AA Terminal-Bench 2.1 | 80.1% | estimated ± 2.5 pp, medium confidence |
| ApprenticeBench | 9.3% | estimated ± 4.0 pp, high confidence |
| ARC-AGI-1 | 92.6% | estimated ± 2.5 pp, high confidence |
| ARC-AGI-2 | 69.6% | estimated ± 9.7 pp, medium confidence |
| CursorBench 3.1 | 58.3% | estimated ± 5.6 pp, low confidence |
| CursorBench 3.2 | 64.2% | estimated ± 3.6 pp, high confidence |
| EBR-bench | 35.2% | estimated ± 11.3 pp, low confidence |
| Terminal-Bench 3.0 | 3.5% | estimated ± 6.9 pp, low confidence |
| FrontierCode 1.1 Main | 46.7% | estimated ± 3.5 pp, low confidence |
| FrontierSWE v2 | 25.4% | estimated ± 8.3 pp, medium confidence |
| Furniture Assembly | 38.9% | estimated ± 12.4 pp, low confidence |
| MaxxBench | 52.6% | estimated ± 5.7 pp, medium confidence |
| MazeBench | 2.7% | estimated ± 4.0 pp, high confidence |
| Mystery Game Puzzles | 30.2% | estimated ± 13.3 pp, low confidence |
| NYT Connections | 68.2% | measured |
| OfficeQA Pro | 66.2% | measured |
| OSWorld-Verified | 78.6% | estimated ± 4.7 pp, high confidence |
| ParseBench-1 | 71.9% | estimated ± 5.4 pp, low confidence |
| Terminal-Bench 4.1 | 22.3% | estimated ± 10.9 pp, low confidence |
| Terminal-Bench Hard | 46.1% | estimated ± 3.9 pp, high confidence |
| Vals SWE-bench | 92.5% | estimated ± 5.2 pp, medium confidence |
| Terminal-Bench 2.1 (Vals AI) | 55.1% | measured |
| CWE-Bench-AA | 60.8% | estimated ± 4.0 pp, medium confidence |
| AA Cyber Index | 32.8% | estimated ± 9.4 pp, medium confidence |
| DeepsecBench-AA | 19.2% | estimated ± 5.2 pp, low confidence |
| HealthBench Professional (length-adjusted, Artificial Analysis) | 57.9% | estimated ± 4.4 pp, low confidence |
| HealthBench Professional (raw, Artificial Analysis) | 63.3% | estimated ± 0.8 pp, low confidence |
| AA ITBench | 47.4% | estimated ± 4.4 pp, high confidence |
| Bug Hunt Bench | 5.7% | estimated ± 3.2 pp, low confidence |
| Chartography (tools) | 86.0% | estimated ± 4.3 pp, low confidence |
| CursorBench 4.0 | 39.9% | estimated ± 3.3 pp, high confidence |
| CWE-bench v1 | 53.0% | measured |
| DeepSWE | 64.3% | measured |
| DRACO | 77.2% | measured |
| FrontierCode 2.0 | 23.4% | estimated ± 6.7 pp, medium confidence |
| FrontierPhysics | 2.5% | estimated ± 4.7 pp, high confidence |
| HLE w/o tools | 43.4% | measured |
| OfficeQA | 77.4% | estimated ± 0.2 pp, medium confidence |
| OSWorld 2.0 | 19.4% | estimated ± 5.5 pp, low confidence |
| OSWorld 2.1 | 69.4% | estimated ± 7.2 pp, low confidence |
| OSWorld 2.1 Offline | 65.7% | estimated ± 5.5 pp, low confidence |
| PostTrainBench v1.1 | 29.3% | estimated ± 4.4 pp, high confidence |
| ProgramBench | 17.5% | measured |
| SpireClimb | 5.8% | estimated ± 7.2 pp, low confidence |
| StealthMark | 43.9% | measured |
| SWE Multilingual | 82.9% | measured |
| SWE-bench Multimodal | 52.2% | estimated ± 8.8 pp, low confidence |
| Toolathlon-Verified | 74.1% | measured |
| VulcanBench Frontier v4 | 88.5% | estimated ± 1.2 pp, low confidence |
| AA AutomationBench | 41.9% | estimated ± 7.9 pp, low confidence |
| GDP.pdf | 19.0% | estimated ± 4.1 pp, medium confidence |
| AA Terminal-Bench 4.0 | 17.7% | estimated ± 11.0 pp, low confidence |
| ARC-AGI-3 | 5.1% | estimated ± 7.2 pp, medium confidence |
| CWE-bench | 47.3% | measured |
| MKA | 39.8% | estimated ± 3.2 pp, high confidence |
| SkillsBench official leaderboard (BenchFlow) | 52.3% | estimated ± 6.1 pp, medium confidence |
| BinaryAudit | 35.8% | estimated ± 12.1 pp, low confidence |
| OTelBench | 22.5% | estimated ± 6.6 pp, medium confidence |
| FACTS Multimodal | 38.9% | estimated ± 4.0 pp, high confidence |
| Hermes Bench | 70.8% | measured |
| Hermes Index | 41.0% | measured |
| CharXiv | 91.7% | estimated ± 5.1 pp, low confidence |
| ExploitBench | 33.9% | estimated ± 12.6 pp, low confidence |
| ExploitGym | 16.4% | estimated ± 5.5 pp, low confidence |
| C-Eval | 93.2% | estimated ± 1.2 pp, medium confidence |
| Claw-Eval | 77.0% | estimated ± 4.5 pp, medium confidence |
| DeepPlanning | 51.8% | estimated ± 13.1 pp, low confidence |
| INCLUDE | 87.2% | estimated ± 2.0 pp, medium confidence |
| MCP Atlas | 83.7% | measured |
| MMLU-ProX | 88.0% | estimated ± 1.3 pp, medium confidence |
| MMLU-Redux | 95.1% | estimated ± 1.7 pp, high confidence |
| NL2Repo | 58.9% | measured |
| PostTrainBench v1 | 31.2% | estimated ± 4.6 pp, high confidence |
| QwenClawBench | 69.0% | estimated ± 2.6 pp, medium confidence |
| SuperGPQA | 94.3% | estimated ± 7.2 pp, low confidence |
| Terminal-Bench 2.0 | 62.6% | estimated ± 5.8 pp, medium confidence |
| Toolathlon | 52.5% | estimated ± 5.2 pp, low confidence |
| HealthBench Hard | 31.2% | estimated ± 5.1 pp, low confidence |
| LiveCodeBench Pro | 88.0% | estimated ± 6.0 pp, low confidence |
| MedXpertQA (Text) | 61.3% | estimated ± 4.5 pp, medium confidence |
| MMLU-Pro (Arcee) | 87.6% | estimated ± 3.4 pp, low confidence |
| ResearchClawBench | 19.7% | estimated ± 1.8 pp, medium confidence |
| PostTrainBench | 27.0% | estimated ± 3.7 pp, high confidence |
| DAYJOB | 2.9% | estimated ± 3.2 pp, high confidence |
| DAYJOB Healthcare | 1.6% | estimated ± 4.7 pp, high confidence |
| BioMysteryBench (human-solvable) | 88.4% | estimated ± 0.6 pp, medium confidence |
| FrontierCode 1.1 Extended | 57.8% | estimated ± 3.0 pp, medium confidence |
| HealthBench Professional | 52.6% | estimated ± 4.2 pp, medium confidence |
| ScarfBench | 64.3% | estimated ± 3.8 pp, medium confidence |
| BlenderWorld (text-only) | 61.0% | estimated ± 4.3 pp, low confidence |
| Cookie-Bench | 75.2% | estimated ± 4.8 pp, low confidence |
| STEMBench | 2.6% | measured |
| SimpleQA | 38.6% | estimated ± 13.4 pp, low confidence |
| Agents' Last Exam | 22.8% | measured |
| OpenHarmony Bench v1.0 | 59.1% | estimated ± 2.7 pp, high confidence |
| SEC-Bench Pro | 51.6% | estimated ± 9.3 pp, medium confidence |
| SWE-Marathon | 31.9% | measured |
| BioBasicsBench | 29.8% | estimated ± 2.2 pp, low confidence |
| ExploitBench Internal Port | 0.0% | estimated ± 6.8 pp, low confidence |
| SWE-Atlas Codebase QnA | 58.6% | estimated ± 4.8 pp, high confidence |