Official benchmarks map / Toolathlon
XiaomiMiMo/MiMo-V2.6-Pro-RL is ranked first on the Toolathlon leaderboard with 76.9, 0.5% ahead of moonshotai/Kimi-K3. Read live from the Hugging Face API on November 6, 2026.
| # | Model | Organization | Score |
|---|---|---|---|
| 1 | XiaomiMiMo/MiMo-V2.6-Pro-RL | XiaomiMiMo | 76.9 |
| 2 | moonshotai/Kimi-K3 | moonshotai | 76.5 |
| 3 | mindlab-research/Macaron-V1.1 | mindlab-research | 76 |
| 4 | deepseek-ai/DeepSeek-V4-Flash-Vision-Exp | deepseek-ai | 75.9 |
| 5 | deepseek-ai/DeepSeek-V4-Pro-0813 | deepseek-ai | 74.1 |
| 6 | tencent/Hy4-preview | tencent | 74.1 |
| 7 | XiaomiMiMo/MiMo-V2.6-Flash-RL | XiaomiMiMo | 73.6 |
| 8 | Qwen/Qwen3.8-Flash-Next | Qwen | 73.5 |
| 9 | zai-org/GLM-5.3 | zai-org | 73 |
| 10 | meta-models/Muse-Spark-Open-816B-A42B | meta-models | 72.8 |
Toolathlon 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/hkust-nlp/Toolathlon
Part of the Hugging Face official benchmarks map by QuantID. Unofficial. Not affiliated with or endorsed by Hugging Face.