Official benchmarks map / deep-swe
PCL-RSI-lab/Seshat-agent is ranked first on the deep-swe leaderboard with 90.3, 17.3% ahead of deepseek-ai/DeepSeek-V4.1-Flash-1017. Read live from the Hugging Face API on November 8, 2026.
| # | Model | Organization | Score |
|---|---|---|---|
| 1 | PCL-RSI-lab/Seshat-agent | PCL-RSI-lab | 90.3 |
| 2 | deepseek-ai/DeepSeek-V4.1-Flash-1017 | deepseek-ai | 77 |
| 3 | moonshotai/Kimi-K3.1-Pro | moonshotai | 75.5 |
| 4 | deepseek-ai/DeepSeek-V4.1-Flash | deepseek-ai | 74.2 |
| 5 | MiniMaxAI/MiniMax-M3.1-Flash | MiniMaxAI | 73.7 |
| 6 | XiaomiMiMo/MiMo-V2.6-Pro-RL | XiaomiMiMo | 71.9 |
| 7 | mindlab-research/Macaron-V1.1 | mindlab-research | 71.7 |
| 8 | XiaomiMiMo/MiMo-V2.6-Flash-RL | XiaomiMiMo | 67.9 |
| 9 | stepfun-ai/Step-5-Preview | stepfun-ai | 67.7 |
| 10 | moonshotai/Kimi-K3 | moonshotai | 67.5 |
deep-swe 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/datacurve/deep-swe
Part of the Hugging Face official benchmarks map by QuantID. Unofficial. Not affiliated with or endorsed by Hugging Face.