OpenAI·released 2024-09-121 source
o1-preview benchmark scores: 8 benchmarks tracked. 38% of its scores are independently reproduced — the rest are self-reported or unverified.
Leads on 0 of 8 benchmarks·best result 81.65 on MATH level 5 (reproduced)·3 of 8 independently reproduced
| Benchmark | Score | Evidence status | Measured |
|---|---|---|---|
| MATH level 5 | 81.65accuracy (%) | reproduced· optimizedT1 | 2024-09-12Epoch AI |
| WeirdML | 47.56accuracy (%) | unverifiedT1 | 2024-09-12https://htihle.github.io/weirdml.html |
| METR Time Horizons | 45.14accuracy (%) | self-reportedT1 | 2024-09-12https://metr.org/blog/2025-03-19-measuring-ai-ability-to-complete-long-tasks/ |
| GPQA diamond | 33.75accuracy (%) | reproduced· optimizedT1 | 2024-09-12Epoch AI |
| OTIS Mock AIME 2024-2025 | 31.04accuracy (%) | reproduced· optimizedT1 | 2024-09-12Epoch AI |
| SimpleBench | 30.04accuracy (%) | unverifiedT1 | 2024-09-12SimpleBench Leaderboard |
| ARC-AGI | 18accuracy (%) | unverified· optimizedT1 | 2024-09-12ARC Prize Leaderboard |
| Cybench | 10accuracy (%) | unverified· optimizedT1 | 2024-09-12Cybench leaderboard |
Claims drawn from cited facts, not live model generation.
Developed by OpenAI, this model was released in 2024, and its vintage indicates a mid-2020s modern foundation-model design.
2 cited facts
This model is scored on 8 tracked benchmarks. Of those, it holds no current top score. Because the results are independently reproduced, outside evaluation supports the record and it can be treated as verified.
3 cited facts
With a reported average gap of 54.33 points behind the leader, this model trails by a wide margin and sits well behind the front of the pack. It tops none of the tracked benchmarks, so it currently holds no category leadership.
2 cited facts
1 facts cross-checked across data sources: 1 single-source.
Source facts, citations, and refresh stamp for this record.
Sources: epoch_benchmarks
o1-preview is an AI model developed by OpenAI, released 2024-09-12. Its benchmark record below tags every score as independently reproduced, self-reported, or unverified.
o1-preview has recorded scores on 8 benchmarks, each shown with its evidence status.
o1-preview has recorded scores on 8 benchmarks — GPQA diamond, MATH level 5, OTIS Mock AIME 2024-2025, WeirdML, Cybench, SimpleBench, and 2 more. The full table above shows each score with its evidence status.
3 of 8 recorded scores (38%) are independently reproduced rather than self-reported by the lab.
o1-preview has 8 tracked claims: 3 independently reproduced, 1 self-reported, 4 unverified.