Web Agent QA on WebWalkerQA
11.5F1 (Easy)ReAct + GraphPO
Evaluation Results
| Method | Links | ||||
|---|---|---|---|---|---|
| ReAct + GraphPOBase Model=Qwen2.5-7B, Agent Framework=ReAct2026.06 | 11.5 | 13.7 | 12.3 | 12.7 | |
| ReAct + GRPOModel Scale=14b, Model Family=Qwen2.52025.09 | 11.4 | 14.8 | 10.3 | 12.4 | |
| ReAct + Tree-GRPOModel Scale=14b, Model Family=Qwen2.52025.09 | 11.1 | 15.5 | 10.8 | 12.8 | |
| ReAct + Tree-GRPOBase Model=Qwen2.5-7B, Agent Framework=ReAct2026.06 | 10.7 | 12.3 | 11.6 | 11.7 | |
| ReActModel Scale=14b, Model Family=Qwen2.52025.09 | 9.5 | 11.3 | 7.4 | 9.5 | |
| ReAct + DAPOBase Model=Qwen2.5-7B, Agent Framework=ReAct2026.06 | 9.5 | 11.7 | 11.4 | 11.3 | |
| DeepSeek-R1-Distill-32bModel Scale=32b, Model Family=DeepSeek, Mode=Direct2025.09 | 9.4 | 13.3 | 9.4 | 11 | |
| ReAct + Tree-GRPOModel Scale=7b, Model Family=Qwen2.52025.09 | 9.3 | 11.8 | 11.9 | 11.2 | |
| ReAct + GRPOModel Scale=7b, Model Family=Qwen2.52025.09 | 8.9 | 11.4 | 11.6 | 10.9 | |
| ReAct + GRPOBase Model=Qwen2.5-7B, Agent Framework=ReAct2026.06 | 8.7 | 11.5 | 10.7 | 10.6 | |
| ReActModel Scale=7b, Model Family=Qwen2.52025.09 | 8 | 9.2 | 5.6 | 7.6 | |
| ReActBase Model=Qwen2.5-7B, Agent Framework=ReAct2026.06 | 7.7 | 9.4 | 5.5 | 7.8 | |
| Qwen2.5-32b-InstructModel Scale=32b, Model Family=Qwen2.5, Mode=Direct2025.09 | 6.2 | 9.4 | 5.8 | 7.4 |