Official benchmarks map / SWE-bench_Verified
ornith-ai/Ornith-1.5-397B is ranked first on the SWE-bench_Verified leaderboard with 86, 0.5% ahead of mindlab-research/Macaron-V1-Venti. Read live from the Hugging Face API on November 8, 2026.
| # | Model | Organization | Score |
|---|---|---|---|
| 1 | ornith-ai/Ornith-1.5-397B | ornith-ai | 86 |
| 2 | mindlab-research/Macaron-V1-Venti | mindlab-research | 85.6 |
| 3 | mindlab-research/Macaron-V1-Coding-Venti | mindlab-research | 85.6 |
| 4 | ornith-ai/Ornith-1.0-397B | ornith-ai | 82.4 |
| 5 | reflection/Beam-1.0-501B-A23B | reflection | 81.6 |
| 6 | deepseek-ai/DeepSeek-V4-Pro | deepseek-ai | 80.6 |
| 7 | MiniMaxAI/MiniMax-M3 | MiniMaxAI | 80.5 |
| 8 | moonshotai/Kimi-K2.6 | moonshotai | 80.2 |
| 9 | thinkingmachines/Inkling-Small | thinkingmachines | 80.2 |
| 10 | deepseek-ai/DeepSeek-V4-Flash | deepseek-ai | 79 |
SWE-bench_Verified 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/SWE-bench/SWE-bench_Verified
Part of the Hugging Face official benchmarks map by QuantID. Unofficial. Not affiliated with or endorsed by Hugging Face.