Rankings / OpenAI
GPT 5 Pro (2025-10-06)
released 2025-10-06
BenchAtlas Index
as of 2026-07-14
base configuration
rank #62
4 families · 3 categories · medium
Benchmark evidence
16 results
external indices
| ECI | medium effort | 150.4 points | independent | Epoch AI Benchmarking Hub |
| ECI | 150.4 points | independent | Epoch AI Benchmarking Hub | |
| ECI | high effort | 150.4 points | independent | Epoch AI Benchmarking Hub |
factuality
| MASK contamination=Potential contamination warning: This model was evaluated after the public release of MASK, allowing model builder access to the prompts and solutions. | 86.0 % 84.4–87.6 | independent | Scale Labs |
knowledge science
| Humanity's Last Exam contamination=Potential contamination warning: This model was evaluated after the public release of HLE, allowing model builder access to the prompts and solutions. · implementation=scale | 31.6 % 29.8–33.5 | independent | Scale Labs |
multimodal
| VISTA | 52.4 % 51.3–53.5 | independent | Scale Labs |
reasoning math
| ARC-AGI-1older version split=public_eval · model_type=CoT | 77.0 % | independent | ARC Prize Leaderboard | |
| ARC-AGI-1older version split=semi_private · model_type=CoT | 70.2 % | independent | ARC Prize Leaderboard | |
| ARC-AGI-2 split=semi_private · model_type=CoT | 18.3 % | independent | ARC Prize Leaderboard | |
| ARC-AGI-2 split=public_eval · model_type=CoT | 13.3 % | independent | ARC Prize Leaderboard | |
| ARC-AGI-2 split=public_eval · model_type=CoT | 8.0 usd_per_task | independent | ARC Prize Leaderboard | |
| ARC-AGI-2 split=semi_private · model_type=CoT | 7.1 usd_per_task | independent | ARC Prize Leaderboard | |
| ARC-AGI-1older version split=semi_private · model_type=CoT | 4.8 usd_per_task | independent | ARC Prize Leaderboard | |
| ARC-AGI-1older version split=public_eval · model_type=CoT | 4.0 usd_per_task | independent | ARC Prize Leaderboard | |
| EnigmaEval contamination=Potential contamination warning: This model was evaluated after the public release of EnigmaEval, allowing model builder access to the prompts and solutions. | 18.8 % 16.5–21.0 | independent | Scale Labs | |
| FrontierMath Tier 4older version | high effort | 14.6 % | independent | Epoch AI Benchmarking Hub 2025-10-09 |