Tool-use task completion on SkillCraft Hard
20Success RateMinimax-M2.1
Evaluation Results
| Method | Links | |
|---|---|---|
| Minimax-M2.1Mode=Skill2026.02 | 20 | |
| Claude 4.5 SonnetMode=Base2026.02 | 20 | |
| Claude 4.5 SonnetMode=Skill2026.02 | 20 | |
| Minimax-M2.1Mode=Base2026.02 | 18 | |
| GPT-5.2Mode=Skill2026.02 | 17 | |
| Gemini 3 ProMode=Skill2026.02 | 17 | |
| GPT-5.2Mode=Base2026.02 | 16 | |
| Gemini 3 ProMode=Base2026.02 | 16 | |
| DeepSeek-V3.2-EXPMode=Skill2026.02 | 15 | |
| DeepSeek-R1Mode=Skill2026.02 | 15 | |
| GLM-4.7Mode=Skill2026.02 | 15 | |
| GLM-4.7Mode=Base2026.02 | 12 | |
| DeepSeek-R1Mode=Base2026.02 | 11 | |
| DeepSeek-V3.2-EXPMode=Base2026.02 | 9 | |
| Kimi-K2-ThinkingMode=Base2026.02 | 8 | |
| Kimi-K2-ThinkingMode=Skill2026.02 | 7 |