Official benchmarks map / gpqa
FINAL-Bench/Darwin-180B-RSI is ranked first on the gpqa leaderboard with 94.44, 0.3% ahead of moonshotai/Kimi-K3.1-Pro. Read live from the Hugging Face API on November 8, 2026.
| # | Model | Organization | Score |
|---|---|---|---|
| 1 | FINAL-Bench/Darwin-180B-RSI | FINAL-Bench | 94.44 |
| 2 | moonshotai/Kimi-K3.1-Pro | moonshotai | 94.2 |
| 3 | moonshotai/Kimi-K3 | moonshotai | 93.5 |
| 4 | FINAL-Bench/Darwin-397B-ZTC | FINAL-Bench | 93.43 |
| 5 | ornith-ai/Ornith-1.5-397B | ornith-ai | 92.8 |
| 6 | Qwen/Qwen3.8-2.4T-A95B | Qwen | 92.6 |
| 7 | tencent/Hy4-preview | tencent | 92.3 |
| 8 | Qwen/Qwen3.8-Flash-Next | Qwen | 91.7 |
| 9 | zai-org/GLM-5.2 | zai-org | 91.2 |
| 10 | FINAL-Bench/Darwin-398B-JGOS | FINAL-Bench | 90.9 |
gpqa is a dataset that Hugging Face marks as an official benchmark. Its leaderboard is built automatically from the .eval_results files that model repositories publish, so entries are mostly self-reported by model authors and evaluation settings differ between submissions. A verified tag appears only where Hugging Face marks an entry as verified.
Original board: huggingface.co/datasets/Idavidrein/gpqa
Part of the Hugging Face official benchmarks map by QuantID. Unofficial. Not affiliated with or endorsed by Hugging Face.