Official benchmarks map / terminal-bench-2.0
mindlab-research/Macaron-V1-Venti is ranked first on the terminal-bench-2.0 leaderboard with 87.6, 27.0% ahead of zai-org/GLM-5.1. Read live from the Hugging Face API on November 6, 2026.
| # | Model | Organization | Score |
|---|---|---|---|
| 1 | mindlab-research/Macaron-V1-Venti | mindlab-research | 87.6 |
| 2 | zai-org/GLM-5.1 | zai-org | 69 |
| 3 | INCModel3/GLM-5.1-MXFP4-Mixed-CT-AutoRound | INCModel3 | 69 |
| 4 | XiaomiMiMo/MiMo-V2.5-Pro | XiaomiMiMo | 68.4 |
| 5 | deepseek-ai/DeepSeek-V4-Pro | deepseek-ai | 67.9 |
| 6 | moonshotai/Kimi-K2.6 | moonshotai | 66.7 |
| 7 | XiaomiMiMo/MiMo-V2.5 | XiaomiMiMo | 65.8 |
| 8 | zai-org/GLM-5.1 | zai-org | 63.5 |
| 9 | INCModel3/GLM-5.1-MXFP4-Mixed-CT-AutoRound | INCModel3 | 63.5 |
| 10 | Qwen/Qwen3.6-27B | Qwen | 59.3 |
terminal-bench-2.0 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/harborframework/terminal-bench-2.0
Part of the Hugging Face official benchmarks map by QuantID. Unofficial. Not affiliated with or endorsed by Hugging Face.