OpenAI · model
GPT-5.3 Codex benchmark scores
As of 2026-11-22, GPT-5.3 Codex (OpenAI) has measured scores on 19 benchmarks and estimated scores on 120 more.
| Benchmark | Score | Source |
|---|---|---|
| AA-GPQA Diamond | 91.5% | measured |
| AA-HLE | 42.5% | measured |
| AA-SciCode | 48.6% | estimated ± 3.6 pp, high confidence |
| Artificial Analysis Intelligence Index | 36.1% | estimated ± 7.2 pp, medium confidence |
| CritPt | 16.9% | measured |
| GDPval-AA | 36.4% | estimated ± 6.2 pp, medium confidence |
| GPQA Diamond | 92.4% | estimated ± 2.4 pp, high confidence |
| HLE | 48.2% | estimated ± 8.5 pp, medium confidence |
| IFBench | 73.6% | estimated ± 6.6 pp, medium confidence |
| AA-LCR | 83.3% | measured |
| LiveCodeBench v6 | 89.7% | estimated ± 2.2 pp, low confidence |
| AA-Omniscience Accuracy | 43.2% | estimated ± 7.7 pp, medium confidence |
| SciCode | 51.3% | estimated ± 2.9 pp, medium confidence |
| BrowseComp | 63.6% | estimated ± 4.7 pp, high confidence |
| HLE w/ tools | 39.9% | estimated ± 5.5 pp, medium confidence |
| IFEval | 93.0% | estimated ± 2.8 pp, high confidence |
| LongBench v2 | 68.6% | estimated ± 5.5 pp, medium confidence |
| τ²-bench results | 86.0% | measured |
| AA Coding Index | 69.4% | estimated ± 5.3 pp, medium confidence |
| AA-MMMU-Pro | 78.5% | measured |
| APEX-Agents-AA | 31.6% | estimated ± 3.6 pp, medium confidence |
| BioMysteryBench (human-difficult) | 47.2% | estimated ± 6.8 pp, low confidence |
| DeepSearchQA | 75.4% | estimated ± 7.9 pp, low confidence |
| SWE-bench Verified | 85.0% | measured |
| BFCL v4 | 64.5% | estimated ± 8.5 pp, low confidence |
| CyberGym | 76.3% | estimated ± 11.0 pp, low confidence |
| JobBench | 33.7% | measured |
| SWE-bench Pro | 56.8% | measured |
| WideResearch | 77.7% | estimated ± 1.2 pp, medium confidence |
| AA Agentic Index | 15.6% | estimated ± 8.0 pp, medium confidence |
| MMLU-Pro | 85.1% | estimated ± 7.0 pp, medium confidence |
| AA-IFBench | 75.4% | measured |
| FailureSensorIQ | 67.6% | estimated ± 1.8 pp, medium confidence |
| GPQA | 91.8% | estimated ± 6.1 pp, medium confidence |
| MMLU | 88.2% | estimated ± 8.3 pp, low confidence |
| Vals GPQA Diamond | 90.6% | estimated ± 3.2 pp, high confidence |
| Vals LiveCodeBench | 87.3% | measured |
| Vals MMLU-Pro | 87.7% | estimated ± 3.0 pp, high confidence |
| MMMLU | 90.5% | estimated ± 0.7 pp, low confidence |
| MMMU | 85.6% | estimated ± 2.7 pp, high confidence |
| AA MMLU-Pro | 88.9% | estimated ± 1.3 pp, medium confidence |
| Gert Labs | 52.0% | measured |
| AA EnterpriseOps-Gym | 44.7% | estimated ± 1.2 pp, medium confidence |
| ApprenticeBench | 4.7% | estimated ± 6.0 pp, low confidence |
| ARC-AGI-1 | 87.4% | estimated ± 7.8 pp, low confidence |
| ARC-AGI-2 | 63.1% | estimated ± 14.4 pp, low confidence |
| CursorBench 3.1 | 52.5% | estimated ± 5.6 pp, low confidence |
| CursorBench 3.2 | 53.5% | estimated ± 3.6 pp, high confidence |
| FrontierCode 1.1 Main | 36.3% | estimated ± 4.2 pp, medium confidence |
| FrontierSWE v2 | 10.1% | estimated ± 8.3 pp, medium confidence |
| MazeBench | 0.9% | estimated ± 2.7 pp, high confidence |
| OfficeQA Pro | 50.7% | estimated ± 6.8 pp, medium confidence |
| OSWorld-Verified | 64.7% | measured |
| ParseBench-1 | 63.4% | estimated ± 6.2 pp, medium confidence |
| Terminal-Bench Hard | 55.3% | estimated ± 4.0 pp, high confidence |
| Vals SWE-bench | 78.0% | measured |
| Terminal-Bench 2.1 (Vals AI) | 68.8% | estimated ± 6.4 pp, medium confidence |
| VulcanBench v3 | 85.2% | estimated ± 5.6 pp, low confidence |
| CWE-Bench-AA | 52.9% | estimated ± 8.1 pp, medium confidence |
| AA Cyber Index | 40.5% | estimated ± 12.4 pp, low confidence |
| DeepsecBench-AA | 18.3% | estimated ± 8.2 pp, medium confidence |
| HealthBench Professional (length-adjusted, Artificial Analysis) | 56.0% | estimated ± 7.3 pp, low confidence |
| HealthBench Professional (raw, Artificial Analysis) | 63.3% | estimated ± 0.8 pp, low confidence |
| AA ITBench | 45.2% | estimated ± 4.6 pp, medium confidence |
| Bug Hunt Bench | 29.8% | estimated ± 12.1 pp, low confidence |
| Chartography (tools) | 82.5% | estimated ± 4.1 pp, low confidence |
| CursorBench 4.0 | 33.2% | estimated ± 3.3 pp, medium confidence |
| CWE-bench v1 | 51.4% | estimated ± 5.7 pp, medium confidence |
| DeepSWE | 56.3% | estimated ± 7.2 pp, low confidence |
| FrontierCode 2.0 | 20.3% | estimated ± 6.7 pp, low confidence |
| HLE w/o tools | 42.3% | estimated ± 5.4 pp, medium confidence |
| OSWorld 2.0 | 0.0% | estimated ± 3.6 pp, medium confidence |
| OSWorld 2.1 | 65.6% | estimated ± 3.6 pp, high confidence |
| OSWorld 2.1 Offline | 60.4% | estimated ± 10.1 pp, low confidence |
| PostTrainBench v1.1 | 14.2% | estimated ± 1.3 pp, low confidence |
| Toolathlon-Verified | 73.3% | estimated ± 3.0 pp, high confidence |
| VulcanBench Frontier v4 | 77.8% | estimated ± 8.3 pp, low confidence |
| AA AutomationBench | 55.3% | estimated ± 9.9 pp, low confidence |
| GDP.pdf | 16.4% | estimated ± 4.1 pp, medium confidence |
| ARC-AGI-3 | 1.3% | estimated ± 9.8 pp, medium confidence |
| ARC-AGI-3 Provider Adapter | 2.9% | estimated ± 7.3 pp, medium confidence |
| CWE-bench | 48.2% | estimated ± 6.2 pp, medium confidence |
| MKA | 36.4% | estimated ± 4.6 pp, high confidence |
| SkillsBench official leaderboard (BenchFlow) | 52.8% | estimated ± 9.5 pp, low confidence |
| BinaryAudit | 42.1% | measured |
| OTelBench | 20.1% | estimated ± 5.6 pp, medium confidence |
| Vibe Code Bench | 61.8% | measured |
| Hermes Index | 33.8% | estimated ± 5.7 pp, medium confidence |
| ParseBench-2 | 69.5% | estimated ± 6.2 pp, medium confidence |
| CharXiv | 83.1% | estimated ± 6.3 pp, low confidence |
| CharXiv w/o tools | 82.4% | estimated ± 2.5 pp, high confidence |
| ExploitGym | 7.4% | estimated ± 8.3 pp, low confidence |
| AI2D_TEST | 94.0% | estimated ± 2.0 pp, medium confidence |
| C-Eval | 94.0% | estimated ± 1.2 pp, low confidence |
| Claw-Eval | 63.0% | estimated ± 4.5 pp, medium confidence |
| CountBench | 97.2% | estimated ± 2.9 pp, medium confidence |
| DynaMath | 85.5% | estimated ± 3.0 pp, medium confidence |
| ERQA | 68.3% | estimated ± 3.1 pp, high confidence |
| MathVision | 89.2% | estimated ± 3.1 pp, high confidence |
| MCP Atlas | 63.0% | estimated ± 8.6 pp, medium confidence |
| MMLU-Redux | 95.4% | estimated ± 3.3 pp, medium confidence |
| MMMU-Pro | 79.3% | estimated ± 1.5 pp, high confidence |
| NL2Repo | 44.2% | estimated ± 4.7 pp, high confidence |
| PostTrainBench v1 | 17.8% | measured |
| RealWorldQA | 85.2% | estimated ± 2.9 pp, medium confidence |
| ScreenSpot Pro | 76.6% | estimated ± 10.8 pp, low confidence |
| Terminal-Bench 2.0 | 77.3% | measured |
| Toolathlon | 45.3% | estimated ± 7.4 pp, low confidence |
| V* | 100.0% | estimated ± 7.5 pp, medium confidence |
| Video-MME (w/o subtitle) | 85.6% | estimated ± 1.7 pp, medium confidence |
| Video-MME (with subtitle) | 87.3% | estimated ± 3.7 pp, medium confidence |
| AA LiveCodeBench | 90.5% | estimated ± 2.6 pp, medium confidence |
| HealthBench Hard | 31.8% | estimated ± 8.6 pp, low confidence |
| LiveCodeBench Pro | 81.2% | estimated ± 6.0 pp, low confidence |
| MedXpertQA (MM) | 76.3% | estimated ± 6.5 pp, low confidence |
| React Native Evals | 81.9% | estimated ± 2.7 pp, medium confidence |
| ResearchClawBench | 15.9% | estimated ± 2.1 pp, medium confidence |
| SWE-Rebench | 58.2% | measured |
| PostTrainBench | 15.4% | estimated ± 1.2 pp, low confidence |
| Finance Agent v2 | 56.0% | estimated ± 2.4 pp, medium confidence |
| BioMysteryBench (human-solvable) | 88.4% | estimated ± 0.6 pp, medium confidence |
| FrontierCode 1.1 Extended | 54.9% | estimated ± 3.0 pp, low confidence |
| HealthBench (raw) | 51.0% | estimated ± 6.0 pp, medium confidence |
| HealthBench (length-adjusted) | 54.5% | estimated ± 3.4 pp, medium confidence |
| HealthBench Professional | 58.2% | estimated ± 4.5 pp, medium confidence |
| HealthBench Professional (raw) | 61.5% | estimated ± 4.5 pp, high confidence |
| HLE-Verified | 38.7% | estimated ± 2.0 pp, medium confidence |
| LABBench2 | 79.4% | estimated ± 1.8 pp, low confidence |
| ScarfBench | 48.9% | estimated ± 12.3 pp, low confidence |
| SimpleQA | 34.4% | estimated ± 7.8 pp, low confidence |
| LiveCodeBench | 82.2% | estimated ± 14.6 pp, low confidence |
| Agents' Last Exam | 31.9% | estimated ± 9.7 pp, medium confidence |
| BabyVision w/ Python | 89.9% | estimated ± 4.0 pp, low confidence |
| OpenHarmony Bench v1.0 | 53.7% | estimated ± 3.2 pp, high confidence |
| MRCRv2 | 85.9% | estimated ± 11.3 pp, low confidence |
| MMLongBench-Doc | 61.5% | estimated ± 2.6 pp, low confidence |
| BioBasicsBench | 34.1% | estimated ± 2.8 pp, medium confidence |
| LifeVal | 60.1% | estimated ± 5.0 pp, medium confidence |
| MMMU-Pro w/ Python | 81.4% | estimated ± 2.0 pp, high confidence |