OpenAI·released 2025-08-04✓ 2 sources
gpt-oss-20b benchmark scores: 7 benchmarks tracked. 0% of its scores are independently reproduced — the rest are self-reported or unverified.
Leads on 0 of 7 benchmarks·best result 65.24 on OTIS Mock AIME 2024-2025 (unverified)·0 of 7 independently reproduced
| Benchmark | Score | Evidence status | Measured |
|---|---|---|---|
| OTIS Mock AIME 2024-2025 | 65.24accuracy (%) | unverifiedT2 | 2025-08-05 |
| GPQA diamond | 47.73accuracy (%) | unverifiedT2 | 2025-08-05 |
| DTBench | 46.67accuracy (%) | unverifiedT1 | 2025-08-05https://conceptualreasoning.ai/dtbench |
| WeirdML | 40.93accuracy (%) | unverifiedT1 | 2025-08-05https://htihle.github.io/weirdml.html |
| LMCA | 17.07accuracy (%) | unverifiedT1 | 2025-08-05https://conceptualreasoning.ai/lmca |
| Terminal Bench | 3.4accuracy (%) | unverified· optimizedT1 | 2025-08-05Terminal-Bench v2 Leaderboard |
| Chess Puzzles | 0accuracy (%) | unverifiedT2 | 2025-08-05 |
Claims drawn from cited facts, not live model generation.
Developed by OpenAI and released in 2025, this model is a mid-size model with 20.91 billion parameters, a scale that balances capability against inference cost. At this scale it offers more headroom than compact models while remaining more efficient and less expensive than frontier-scale systems.
4 cited facts
This model is scored on 7 tracked benchmarks, but it holds no current top score on any of them. Its record remains unverified: no outside evaluation has either confirmed or disputed the reported numbers.
3 cited facts
On average it trails the benchmark leader by 57.65 points, a score gap under the disclosed harness that reads as well behind the leaders rather than a capability claim. It holds the top score on 0 benchmarks, meaning no genuine category leadership is shown yet.
2 cited facts
2 facts cross-checked across data sources: 1 corroborated, 1 single-source.
Source facts, citations, and refresh stamp for this record.
Sources: epoch_benchmarks, huggingface_models
gpt-oss-20b is an AI model developed by OpenAI, released 2025-08-04. Its benchmark record below tags every score as independently reproduced, self-reported, or unverified.
gpt-oss-20b has recorded scores on 7 benchmarks, each shown with its evidence status.
gpt-oss-20b has recorded scores on 7 benchmarks — GPQA diamond, OTIS Mock AIME 2024-2025, DTBench, WeirdML, LMCA, Chess Puzzles, and 1 more. The full table above shows each score with its evidence status.
0 of 7 recorded scores (0%) are independently reproduced rather than self-reported by the lab.
gpt-oss-20b has 7 tracked claims: 7 unverified.