Textual Environment Interaction on Mind2Web
92.44BSQWEN-2.5-3B-SWIRL
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| QWEN-2.5-3B-SWIRLBackbone=Qwen-2.5, Size=3B, Mode=SWIRL2026.02 | 92.44 | 49.1 | |
| QWEN-2.5-3B-SFTBackbone=Qwen-2.5, Size=3B, Mode=SFT2026.02 | 92.37 | 48.97 | |
| QWEN-2.5-3B-INSTRUCTBackbone=Qwen-2.5, Size=3B, Mode=Instruct2026.02 | 82.11 | 18.01 | |
| DEEPSEEK-7B-CHATBackbone=DeepSeek, Size=7B, Mode=Chat2026.02 | 80.18 | 14.32 | |
| QWEN-2.5-32B-INSTRUCTBackbone=Qwen-2.5, Size=32B, Mode=Instruct2026.02 | 79.01 | 11.86 | |
| QWEN-2.5-7B-INSTRUCTBackbone=Qwen-2.5, Size=7B, Mode=Instruct2026.02 | 77.03 | 7.35 | |
| OLMO-3-7B-INSTRUCTBackbone=OLMo-3, Size=7B, Mode=Instruct2026.02 | 76.05 | 6 | |
| QWEN-2.5-14B-INSTRUCTBackbone=Qwen-2.5, Size=14B, Mode=Instruct2026.02 | 75.43 | 3.02 |