Web-based Task Automation on WebShop (test)
85.5ScoreECPO
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| ECPOBase Model=Qwen2.5-7B-Instruct2026.06 | 85.5 | 74.7 | |
| ECPOBase Model=Qwen2.5-1.5B-Instruct2026.06 | 84.4 | 71.9 | |
| GiGPOw/stdBase Model=Qwen2.5-7B-Instruct2026.06 | 84.4 | 72.4 | |
| PPOBase Model=Qwen2.5-7B-Instruct2026.06 | 81.4 | 68.7 | |
| GiGPOw/stdBase Model=Qwen2.5-1.5B-Instruct2026.06 | 80.4 | 64.6 | |
| RLOOBase Model=Qwen2.5-7B-Instruct2026.06 | 80.3 | 65.7 | |
| GRPOBase Model=Qwen2.5-7B-Instruct2026.06 | 79.3 | 66.1 | |
| RLOOBase Model=Qwen2.5-1.5B-Instruct2026.06 | 73.9 | 52.1 | |
| PPOBase Model=Qwen2.5-1.5B-Instruct2026.06 | 73.8 | 51.5 | |
| GRPOBase Model=Qwen2.5-1.5B-Instruct2026.06 | 64.1 | 52.6 | |
| ReflexionBase Model=Qwen2.5-7B-Instruct2026.06 | 58.1 | 28.8 | |
| ReflexionBase Model=Qwen2.5-1.5B-Instruct2026.06 | 55.8 | 21.9 | |
| ReActBase Model=Qwen2.5-7B-Instruct2026.06 | 46.2 | 19.5 | |
| ReActBase Model=Qwen2.5-1.5B-Instruct2026.06 | 40.1 | 11.3 | |
| PromptingBase Model=Qwen2.5-7B-Instruct2026.06 | 26.4 | 7.8 | |
| PromptingBase Model=Qwen2.5-1.5B-Instruct2026.06 | 23.1 | 5.2 | |
| GLM-5.1Model Family=Closed-source LLMs2026.06 | 12.9 | 8.1 | |
| Gemini-3.5-FlashModel Family=Closed-source LLMs2026.06 | 5.8 | 3.9 |