Collaborative Interaction on CollabBench Cook-MultiPlayer
143.47ScoreGPT-5.2
Evaluation results
| Method | Links | ||||||
|---|---|---|---|---|---|---|---|
| GPT-5.2Method Category=Oracle, Agent Role=Agent 12026.06 | 143.47 | 53.02 | 0.3 | 2.68 | 3.84 | 3.47 | |
| GPT-5.2Method Category=Base, Agent Role=Agent 22026.06 | 137.12 | 42.4 | 0.2 | 1.88 | 3.1 | 2.52 | |
| DeepSeek-V3.1Method Category=Base, Agent Role=Agent 22026.06 | 136.8 | 46.24 | 0.31 | 2.07 | 3.04 | 2.77 | |
| DeepSeek-V3.1Method Category=Base, Agent Role=Agent 12026.06 | 136.53 | 40.3 | 0.31 | 1.79 | 2.97 | 2.67 | |
| Qwen2.5-72B-InstructMethod Category=Base, Agent Role=Agent 12026.06 | 135.47 | 41.99 | 0.27 | 1.37 | 2.77 | 2.45 | |
| GPT-5.2Method Category=Base, Agent Role=Agent 12026.06 | 135.2 | 41.96 | 0.2 | 1.63 | 2.89 | 2.27 | |
| GPT-5.2Method Category=Oracle, Agent Role=Agent 22026.06 | 134.9 | 33.56 | 0.2 | 2.34 | 3.45 | 3.46 | |
| Qwen2.5-72B-InstructMethod Category=Base, Agent Role=Agent 22026.06 | 114.13 | 39.45 | 0.26 | 1.26 | 2.65 | 2.48 | |
| Qwen2.5-7B-InstructMethod Category=Trained, Agent Role=Agent 22026.06 | 102.4 | 34.76 | 0.22 | 0.55 | 1.99 | 2.08 | |
| Qwen2.5-7B-InstructMethod Category=Trained, Agent Role=Agent 12026.06 | 99.2 | 34.03 | 0.23 | 0.74 | 2.26 | 2.12 | |
| Qwen2.5-7B-InstructMethod Category=Base, Agent Role=Agent 12026.06 | 86.93 | 33.3 | 0.23 | 0.45 | 1.92 | 1.86 | |
| Qwen2.5-7B-InstructMethod Category=Base, Agent Role=Agent 22026.06 | 85.87 | 33.73 | 0.23 | 0.53 | 1.87 | 1.88 |