OpenAI · model
GPT-5.6 Luna benchmark scores
As of 2026-10-07, GPT-5.6 Luna (OpenAI) has measured scores on 43 benchmarks and estimated scores on 82 more.
| Benchmark | Score | Source |
|---|---|---|
| AA-GPQA Diamond | 91.1% | measured |
| AA-HLE | 39.5% | measured |
| AA-SciCode | 53.6% | measured |
| AIME26 | 100.0% | estimated ± 0.6 pp, low confidence |
| Artificial Analysis Intelligence Index | 51.2% | measured |
| CritPt | 20.6% | measured |
| GDPval-AA | 48.2% | measured |
| GPQA Diamond | 92.3% | measured |
| HLE | 51.6% | estimated ± 5.5 pp, low confidence |
| AA-LCR | 83.7% | measured |
| LiveCodeBench v6 | 90.8% | estimated ± 2.1 pp, high confidence |
| AA-Omniscience Accuracy | 42.7% | measured |
| SciCode | 53.2% | estimated ± 3.2 pp, high confidence |
| Terminal-Bench 2.1 | 84.7% | measured |
| BrowseComp | 83.3% | measured |
| HLE w/ tools | 57.0% | estimated ± 4.6 pp, high confidence |
| LongBench v2 | 61.5% | estimated ± 5.9 pp, low confidence |
| VITA-Bench | 45.3% | estimated ± 9.9 pp, low confidence |
| τ²-bench results | 95.4% | estimated ± 7.3 pp, medium confidence |
| AA Coding Index | 71.5% | measured |
| AA-MMMU-Pro | 78.6% | measured |
| APEX-Agents | 39.9% | estimated ± 10.0 pp, low confidence |
| APEX-Agents-AA | 35.8% | measured |
| BioMysteryBench (human-difficult) | 42.3% | estimated ± 5.9 pp, low confidence |
| DeepSearchQA | 92.2% | estimated ± 4.5 pp, high confidence |
| SWE-bench Verified | 84.4% | estimated ± 2.5 pp, high confidence |
| AutomationBench | 28.4% | estimated ± 5.8 pp, medium confidence |
| BFCL v4 | 77.3% | estimated ± 2.6 pp, medium confidence |
| CyberGym | 77.9% | measured |
| JobBench | 62.5% | estimated ± 4.9 pp, medium confidence |
| SWE-bench Pro | 62.7% | measured |
| τ³-bench results | 71.9% | estimated ± 2.1 pp, low confidence |
| WideResearch | 79.2% | estimated ± 4.6 pp, high confidence |
| AA Agentic Index | 42.7% | measured |
| MMLU-Pro | 86.9% | estimated ± 3.0 pp, high confidence |
| FrontierMath v2 (Tier 4) | 58.5% | measured |
| FrontierMath v2 (Tiers 1-3) | 78.6% | measured |
| GPQA | 92.3% | measured |
| Gert Labs | 64.5% | estimated ± 3.0 pp, medium confidence |
| AA-AnalystAgent | 54.4% | estimated ± 5.5 pp, medium confidence |
| AA EnterpriseOps-Gym | 44.5% | estimated ± 6.7 pp, low confidence |
| AA Harvey LAB | 92.7% | estimated ± 1.2 pp, medium confidence |
| ApprenticeBench | 7.0% | measured |
| ARC-AGI-1 | 88.0% | measured |
| ARC-AGI-2 | 59.5% | measured |
| CursorBench 3.1 | 52.9% | estimated ± 3.4 pp, medium confidence |
| CursorBench 3.2 | 61.1% | measured |
| Terminal-Bench 3.0 | 14.3% | measured |
| FrontierCode 1.1 Main | 43.9% | estimated ± 1.6 pp, medium confidence |
| FrontierSWE v2 | 14.4% | estimated ± 7.6 pp, low confidence |
| OfficeQA Pro | 50.8% | estimated ± 7.2 pp, medium confidence |
| OSWorld-Verified | 74.2% | estimated ± 1.9 pp, medium confidence |
| Vals GPQA Diamond | 91.7% | measured |
| Vals LiveCodeBench | 84.0% | estimated ± 2.7 pp, low confidence |
| Vals MMLU-Pro | 86.0% | measured |
| Vals SWE-bench | 93.0% | measured |
| Terminal-Bench 2.1 (Vals AI) | 79.0% | measured |
| VulcanBench v3 | 85.5% | measured |
| AA AutomationBench | 59.6% | estimated ± 5.5 pp, low confidence |
| GDP.pdf | 22.0% | estimated ± 4.4 pp, low confidence |
| AA ITBench | 46.3% | estimated ± 3.9 pp, low confidence |
| AA Tau3 Banking | 47.4% | estimated ± 5.8 pp, medium confidence |
| AA Terminal-Bench 2.1 | 84.9% | estimated ± 3.5 pp, high confidence |
| AA Terminal-Bench 4.0 | 24.3% | estimated ± 8.9 pp, medium confidence |
| CursorBench 4.0 | 35.9% | measured |
| CWE-bench v1 | 54.6% | estimated ± 6.9 pp, medium confidence |
| DeepSWE | 67.2% | measured |
| HLE w/o tools | 42.1% | estimated ± 4.4 pp, high confidence |
| OSWorld 2.0 | 45.6% | measured |
| PostTrainBench v1.1 | 33.1% | estimated ± 3.3 pp, high confidence |
| ProgramBench | 77.6% | estimated ± 6.0 pp, low confidence |
| Toolathlon-Verified | 78.3% | estimated ± 1.0 pp, medium confidence |
| Vibe Code Bench | 56.5% | estimated ± 10.3 pp, low confidence |
| CharXiv | 81.5% | estimated ± 5.4 pp, medium confidence |
| CharXiv w/o tools | 82.7% | estimated ± 1.8 pp, medium confidence |
| ExploitBench | 33.2% | measured |
| ExploitGym | 12.4% | measured |
| C-Eval | 94.4% | estimated ± 0.9 pp, low confidence |
| Claw-Eval | 68.5% | estimated ± 7.9 pp, medium confidence |
| HMMT Nov 2025 | 93.2% | estimated ± 2.6 pp, low confidence |
| MathVision | 86.9% | estimated ± 3.6 pp, high confidence |
| MCP Atlas | 78.7% | estimated ± 7.8 pp, low confidence |
| MCP-Tasks | 78.3% | estimated ± 7.0 pp, low confidence |
| MMAnswerBench | 95.1% | estimated ± 1.3 pp, low confidence |
| MMLU-Redux | 96.9% | estimated ± 1.1 pp, low confidence |
| MMMU-Pro | 78.4% | measured |
| NL2Repo | 56.7% | estimated ± 4.7 pp, high confidence |
| QwenClawBench | 100.0% | estimated ± 6.0 pp, low confidence |
| ScreenSpot Pro | 74.0% | estimated ± 10.0 pp, low confidence |
| SuperGPQA | 82.5% | estimated ± 7.5 pp, medium confidence |
| Terminal-Bench 2.0 | 79.2% | estimated ± 5.9 pp, low confidence |
| Toolathlon | 53.4% | measured |
| VideoMMMU | 85.1% | estimated ± 1.0 pp, high confidence |
| ERQA | 67.9% | estimated ± 3.9 pp, high confidence |
| HealthBench Hard | 32.0% | measured |
| LiveCodeBench Pro | 85.8% | estimated ± 6.0 pp, low confidence |
| MedXpertQA (MM) | 71.1% | estimated ± 2.4 pp, high confidence |
| MedXpertQA (Text) | 61.3% | estimated ± 4.5 pp, medium confidence |
| MMLU-Pro (Arcee) | 87.6% | estimated ± 3.4 pp, low confidence |
| React Native Evals | 100.0% | estimated ± 2.4 pp, low confidence |
| ResearchClawBench | 20.3% | estimated ± 2.0 pp, medium confidence |
| SWE-Rebench | 58.1% | estimated ± 8.9 pp, low confidence |
| ARC-AGI-3 | 0.2% | measured |
| FrontierMath (legacy) | 78.6% | measured |
| BioMysteryBench (human-solvable) | 89.2% | estimated ± 0.6 pp, medium confidence |
| DRACO | 81.3% | estimated ± 3.0 pp, medium confidence |
| FrontierCode 1.1 Extended | 55.1% | measured |
| HealthBench (raw) | 48.6% | estimated ± 6.4 pp, low confidence |
| HealthBench (length-adjusted) | 53.8% | estimated ± 0.9 pp, medium confidence |
| HealthBench Professional | 55.7% | measured |
| HealthBench Professional (raw) | 59.0% | estimated ± 5.0 pp, medium confidence |
| HLE-Verified | 35.3% | estimated ± 0.7 pp, low confidence |
| LABBench2 | 67.6% | estimated ± 1.3 pp, low confidence |
| MMMU | 86.0% | estimated ± 0.9 pp, medium confidence |
| LiveCodeBench | 100.0% | estimated ± 8.9 pp, low confidence |
| Agents' Last Exam | 24.0% | estimated ± 2.1 pp, high confidence |
| OpenHarmony Bench | 58.7% | estimated ± 2.4 pp, high confidence |
| SEC-Bench Pro | 54.9% | estimated ± 4.3 pp, medium confidence |
| SimpleVQA | 67.2% | estimated ± 8.2 pp, medium confidence |
| MMMLU | 90.6% | estimated ± 1.1 pp, medium confidence |
| MMLU | 93.9% | estimated ± 1.1 pp, low confidence |
| MMMU-Pro w/ Python | 79.5% | measured |
| RealWorldQA | 86.9% | estimated ± 1.2 pp, medium confidence |
| Video-MME (with subtitle) | 87.7% | estimated ± 1.8 pp, medium confidence |
| OmniDocBench 1.5 | 91.7% | estimated ± 5.9 pp, low confidence |