Web-based Agent Reasoning on WebWalkerQA Easy
72.5Pass@3ExpSeek
Evaluation Results
| Method | Links | |
|---|---|---|
| ExpSeekBase Model=Qwen3-32B2026.01 | 72.5 | |
| ExpSeekBase Model=Qwen3-8B2026.01 | 69.17 | |
| REASONINGBANK+Base Model=Qwen3-32B2026.01 | 68.33 | |
| Training-Free GRPOBase Model=Qwen3-32B2026.01 | 63.95 | |
| No ExperienceBase Model=Qwen3-32B2026.01 | 62.5 | |
| REASONINGBANK+Base Model=Qwen3-8B2026.01 | 61.17 | |
| No ExperienceBase Model=Qwen3-8B2026.01 | 57.5 | |
| Training-Free GRPOBase Model=Qwen3-8B2026.01 | 56.67 |