o1 ranks first on 18 of the 145 benchmarks with five or more models scored.
Closed weights Flagship
Its best score on each benchmark, and where that ranks among every model with a score there. We rank each model once, at its best setting.
| Benchmark | Score | Rank | Source | |
|---|---|---|---|---|
| MMLU-Pro | 89.3% | #3 of 212 | Official leaderboard | |
| MMLU | 91.8% | #2 of 131 | Lab model card | |
| AraGen | 84.3% | #1 of 42 | Official leaderboard | |
| SEAL Multilingual | 1,134 | #1 of 19 | Official leaderboard | |
| SEAL Precise Instruction Following | 92.0 | #1 of 18 | Official leaderboard | |
| AidanBench | 3204.0 | #1 of 13 | Official leaderboard | |
| ECLeKTic | 38.7% | #3 of 35 | Official leaderboard | |
| LLM Confabulation Leaderboard | 11.7% | #6 of 58 | Official leaderboard | |
| MARS-Bench | 59.6% | #2 of 15 | Official leaderboard | |
| Fiction.LiveBench | 83.3% | #8 of 57 | Fiction.live leaderboard, via Epoch AI | |
| SuperGPQA | 60.2% | #10 of 59 | Official leaderboard | |
| Vectara Hallucination Leaderboard (HHEM) | 2.40% | #31 of 148 | Official leaderboard | |
| SimpleQA | 47% | #14 of 60 | Lab model card | |
| MGSM | 89.3% | #14 of 57 | Lab model card | |
| BuzzBench | 59.3 | #5 of 20 | Official leaderboard | |
| Internal API instruction following (OpenAI) | 51.3% | #2 of 8 | Official leaderboard | |
| ComplexNLU (Tencent) | 64.7% | #1 of 4 | Official leaderboard | |
| Arena-Hard v2.0 | 61% | #6 of 23 | Official leaderboard | |
| LMArena (Elo) | 1,402 | #127 of 335 | Official leaderboard | |
| AbstentionBench | 74.8% | #2 of 5 | Official leaderboard | |
| FACTS Grounding | 78.8% | #15 of 34 | Official leaderboard | |
| SimpleQA Verified | 41.1% | #52 of 94 | Epoch AI run | |
| PersonQA | 55% | #5 of 9 | Lab model card | |
| SimpleQA Hallucination Rate | 44% | #7 of 12 | Lab model card | |
| Humanity's Last Exam | 7.96% | #35 of 47 | Official leaderboard | |
| HUMAINE | 16 | #48 of 50 | Official leaderboard | |
| HLE Calibration | 83.0 | #4 of 4 | Official leaderboard |
| Benchmark | Score | Rank | Source | |
|---|---|---|---|---|
| U-MATH | 86.8% | #1 of 30 | Official leaderboard | |
| ZebraLogic | 81% | #2 of 46 | Official leaderboard | |
| MATH Level 5 | 94.7% | #13 of 94 | Epoch AI run | |
| 3SAT-Search | 89% | #2 of 11 | Official leaderboard | |
| GSM8K-Platinum | 99.7% | #2 of 8 | Official leaderboard | |
| PolyMath | 38.9% | #12 of 38 | Official leaderboard | |
| GSM8K Platinum Initial 268 | 99.3% | #6 of 19 | Official leaderboard | |
| IneqMath | 8.00% | #11 of 34 | Official leaderboard | |
| QuantBench | 60% | #1 of 3 | Lab model card | |
| TSP-Opt | 45.8% | #4 of 11 | Official leaderboard | |
| MATH-500 | 96.4% | #27 of 70 | Lab model card | |
| GPQA Diamond | 76.8% | #105 of 229 | Epoch AI run | |
| OTIS Mock AIME 2024-2025 | 73.3% | #102 of 212 | Epoch AI run | |
| CombiBench | 2.00% | #2 of 4 | Official leaderboard | |
| AIME 2024 | 79.2% | #45 of 88 | Lab model card | |
| FrontierMath | 9.31% | #32 of 55 | Official leaderboard | |
| DTBench | 74.7% | #108 of 180 | Official leaderboard | |
| AIME 2025 | 81.7% | #33 of 53 | Official leaderboard | |
| SimpleBench | 40.1% | #74 of 106 | Official leaderboard | |
| PHYBench | 18% | #20 of 28 | Official leaderboard | |
| HMMT 2025 | 48.3% | #39 of 53 | Official leaderboard | |
| FrontierMath Tiers 1-3 v2 | 14.7% | #80 of 86 | Official leaderboard | |
| FrontierScience-Olympiad | 42.5% | #7 of 7 | Official leaderboard | |
| FrontierScience-Research | 3.17% | #7 of 7 | Official leaderboard |
| Benchmark | Score | Rank | Source | |
|---|---|---|---|---|
| BigCodeBench-Hard | 35.5% | #2 of 96 | Official leaderboard | |
| HumanEval-V | 40.6% | #1 of 24 | Official leaderboard | |
| Aider-Edit | 84.2% | #2 of 34 | Official leaderboard | |
| OpenAI Research Engineer Interview: Coding | 100% | #1 of 7 | Lab model card | |
| Aider Polyglot | 61.7% | #12 of 46 | Official leaderboard | |
| SEAL Coding | 1,083 | #8 of 26 | Official leaderboard | |
| PaperBench | 26% | #1 of 3 | Official leaderboard | |
| Multi-SWE-bench | 6.43% | #4 of 11 | Official leaderboard | |
| METR Time Horizon | 42 min | #12 of 31 | Official leaderboard | |
| OIBench | 15.4% | #4 of 9 | Official leaderboard | |
| LiveSQLBench | 25.2% | #17 of 36 | Official leaderboard | |
| WeirdML | 46.1% | #57 of 112 | Official leaderboard | |
| ArtifactsBench (Gemini 2.5 Pro 0506 judge) | 38.6% | #10 of 18 | Official leaderboard | |
| OpenAI Research Engineer Interview: Multiple Choice | 78% | #5 of 9 | Lab model card | |
| METR Time Horizon 1.1 | 39 min | #15 of 26 | Official leaderboard | |
| Spider 2.0 | 23.2% | #16 of 23 | Official leaderboard | |
| MLE-bench | 14% | #7 of 10 | Lab model card | |
| SWE-bench Verified | 48.9% | #134 of 155 | Lab model card | |
| TasteVal | 6.71× | #18 of 20 | Official leaderboard | |
| PaperBench Code-Dev | 43.4% | #1 of 1 | Official leaderboard |
| Benchmark | Score | Rank | Source | |
|---|---|---|---|---|
| ToolComp | 70.1% | #1 of 32 | Official leaderboard | |
| C3-Bench | 12.1% | #2 of 30 | Official leaderboard | |
| Step Race | 4 | #5 of 59 | Official leaderboard | |
| CRMArena | 64.3% | #1 of 9 | Official leaderboard | |
| Berkeley Function Calling Leaderboard (BFCL) | 67.8% | #10 of 67 | Official leaderboard | |
| lmgame-Bench | 46.5 | #4 of 23 | Official leaderboard | |
| BrowseComp-ZH | 29.1% | #2 of 9 | Official leaderboard | |
| tau-bench | 63.9% | #8 of 23 | Lab model card | |
| LLM Chess | 280 | #50 of 123 | Official leaderboard | |
| Galileo Agent Leaderboard | 0.88 | #13 of 29 | Official leaderboard | |
| GAIA | 49.8% | #34 of 64 | Official leaderboard | |
| Elimination Game | 4 | #31 of 56 | Official leaderboard | |
| Pencil Puzzle Bench | 0% | #31 of 53 | Official leaderboard | |
| Chess Puzzles | 11% | #83 of 93 | Epoch AI run |
| Benchmark | Score | Rank | Source | |
|---|---|---|---|---|
| MedXpertQA-MM | 56.3% | #1 of 9 | Official leaderboard | |
| MedXpertQA | 49.9% | #1 of 4 | Official leaderboard | |
| TaxEval v2 | 74.3% | #38 of 129 | Official leaderboard | |
| CadEval | 56% | #4 of 13 | Official leaderboard | |
| LegalBench | 80.4% | #83 of 138 | Official leaderboard | |
| HealthBench | 41.8% | #5 of 8 | Official leaderboard | |
| HealthBench Hard | 7.91% | #5 of 8 | Official leaderboard | |
| Vals Finance Agent v1 | 20.8% | #33 of 47 | Official leaderboard |
| Benchmark | Score | Rank | Source | |
|---|---|---|---|---|
| CoreCognition | 74.9% | #1 of 119 | Official leaderboard | |
| MMVU | 75.5% | #1 of 28 | Official leaderboard | |
| CharXiv (DQ) | 88.9% | #4 of 67 | Official leaderboard | |
| OCR-Reasoning | 44.4% | #1 of 15 | Official leaderboard | |
| SpatialViz-Bench | 41.4% | #1 of 14 | Official leaderboard | |
| MMMU | 78.2% | #11 of 136 | Official leaderboard | |
| CSVQA | 49.6% | #1 of 12 | Official leaderboard | |
| CharXiv | 55.1% | #8 of 77 | Official leaderboard | |
| VCR-Bench | 56.7% | #1 of 9 | Official leaderboard | |
| VisualPuzzles | 51.8% | #4 of 29 | Official leaderboard | |
| VHELM | 71.6 | #7 of 41 | Official leaderboard | |
| MicroVQA | 52.8% | #5 of 19 | Official leaderboard | |
| MORSE-500 | 19.8% | #3 of 11 | Official leaderboard | |
| GeoBench | 80% | #15 of 48 | Official leaderboard | |
| MATH-Vision (MathVision) | 60.3% | #35 of 105 | Official leaderboard | |
| VISTA | 45.3% | #27 of 54 | Official leaderboard | |
| Vision Arena | 1,193 | #83 of 136 | Official leaderboard | |
| ViVerBench | 71.5% | #5 of 8 | Official leaderboard | |
| MathVista | 71% | #44 of 70 | Lab model card | |
| VPCT | 37% | #21 of 27 | VPCT leaderboard, via Epoch AI | |
| CaP-X (CaP-Bench) | 8.67% | #8 of 10 | Official leaderboard | |
| VisFactor | 20.8% | #21 of 26 | Official leaderboard | |
| Drone-Bench | 31.9% | #17 of 19 | Official leaderboard | |
| ZeroBench | 0% | #75 of 76 | Official leaderboard |
| Benchmark | Score | Rank | Source | |
|---|---|---|---|---|
| Multimodal Troubleshooting Virology | 59% | #2 of 27 | Lab model card | |
| BaxBench | 48.2% | #4 of 34 | Official leaderboard | |
| HELM Safety | 97.6 | #9 of 75 | Official leaderboard | |
| Standard Refusal Evaluation | 100% | #1 of 8 | Lab model card | |
| ChangeMyView | 89.1% | #1 of 5 | Lab model card | |
| AIR-Bench 2024 | 80% | #23 of 82 | Official leaderboard | |
| Challenging Refusal Evaluation | 92% | #3 of 9 | Lab model card | |
| MakeMeSay | 42% | #4 of 10 | Lab model card | |
| FORTRESS | 19.4% | #30 of 73 | Official leaderboard | |
| Molecular Biology Capabilities Test | 51.1% | #37 of 90 | Official leaderboard | |
| SimpleSafetyTests | 99% | #34 of 75 | Official leaderboard | |
| Gray Swan Agent Red Teaming (ART) benchmark | 2.56% | #4 of 8 | Official leaderboard | |
| MakeMePay | 27% | #3 of 6 | Lab model card | |
| Contextual Nuclear Knowledge | 74% | #2 of 4 | Lab model card | |
| Political Persuasion Parallel Generation | 47.1% | #2 of 4 | Lab model card | |
| Radiological and Nuclear Expert Knowledge | 70% | #2 of 4 | Lab model card | |
| WildChat (toxic conversations) | 97.9% | #2 of 4 | Lab model card | |
| CoT Deception Monitoring | 0.17% | #1 of 2 | Lab model card | |
| Red Teaming Network pairwise safety comparison | 59.8% | #1 of 2 | Lab model card | |
| Tacit Knowledge Brainstorm (Open-Ended) | 51% | #1 of 2 | Lab model card | |
| Virology Capabilities Test (VCT) | 35.4% | #4 of 7 | Official leaderboard | |
| Jailbreak Augmented Examples | 100% | #3 of 5 | Lab model card | |
| Human Pathogen Capabilities Test | 48.2% | #53 of 86 | Official leaderboard | |
| Anthropic Red Team | 98.3% | #50 of 75 | Official leaderboard | |
| BioLP-bench | 33% | #6 of 9 | Lab model card | |
| Capture the Flag (High School) | 46% | #6 of 9 | Lab model card | |
| XSTest | 93.6% | #11 of 16 | Lab model card | |
| Capture the Flag (Professional) | 13% | #16 of 21 | Lab model card | |
| Capture the Flag (Collegiate) | 13% | #10 of 13 | Lab model card | |
| World-Class Bio | 31.2% | #61 of 74 | Official leaderboard | |
| Tacit Knowledge and Troubleshooting | 66% | #25 of 30 | Lab model card | |
| ProtocolQA Open-Ended | 24% | #24 of 27 | Lab model card | |
| StrongREJECT | 72% | #17 of 19 | Lab model card | |
| Human Sourced Jailbreaks | 94% | #8 of 8 | Lab model card | |
| Production Jailbreaks | 99% | #5 of 5 | Lab model card | |
| Image input safety evaluations | 96.1% | #2 of 2 | Lab model card |
| Benchmark | Score | Rank | Source | |
|---|---|---|---|---|
| LiveBench | 76.6% | #5 of 27 | Lab model card | |
| ACCoRD | 64.2 | #77 of 152 | Official leaderboard | |
| Conceptual Reasoning Index | 37.8 | #103 of 151 | Official leaderboard | |
| LMCA | 22.3% | #107 of 147 | Official leaderboard | |
| CAIS Vision Capabilities Index | 40.9 | #43 of 53 | Official leaderboard | |
| CAIS Risk Index | 64.0 | #56 of 67 | Official leaderboard | |
| CAIS Text Capabilities Index | 8.00 | #65 of 68 | Official leaderboard |
The Self-Reported Index fits one score to the 11 capability results OpenAI reported for this model, on the scale of the Epoch Capabilities Index. Labs choose which benchmarks to show and how to run them, so read it as the lab's claim, not an independent measurement. The 90% interval runs from 134.9 to 139.9. How we compute it.
Where the benchmark's own leaderboard, or Epoch AI's run, also has this model. Gaps mostly come from different harnesses, effort levels and task subsets. Every pair.
| Benchmark | Lab says | Board says | Gap |
|---|---|---|---|
| MATH-500 | 96.4% MATH (pass @1) | 90.4% Leaderboard, Vals AI harness | +6.0 pts |
| SimpleQA | 47% SimpleQA, accuracy (Table 3) | 42.6% Leaderboard | +4.4 pts |
| MathVista | 71% MathVista (pass @1) | 73.9% Leaderboard | −2.9 pts |
| GPQA Diamond | 75.7% GPQA diamond | 76.8% Epoch AI run, high | −1.1 pts |
| MMMU | 77.3% MMMU (pass @1) | 78.2% Leaderboard | −0.9 pts |
OpenAI's models in color, against every other model on the Epoch Capabilities Index, by release date.
Source: Epoch Capabilities Index, Epoch AI (CC BY 4.0)
The benchmarks the lab chose to headline, with the scores it gave.
Every result in the longer documents the lab published with it. 45 of them aren't in the launch post.