Web-based Agent Reasoning on WebWalkerQA Medium
72.86Pass@3ExpSeek
Evaluation Results
| Method | Links | |
|---|---|---|
| ExpSeekBase Model=Qwen3-32B2026.01 | 72.86 | |
| ExpSeekBase Model=Qwen3-8B2026.01 | 72.45 | |
| REASONINGBANK+Base Model=Qwen3-32B2026.01 | 64.33 | |
| Training-Free GRPOBase Model=Qwen3-8B2026.01 | 62.54 | |
| No ExperienceBase Model=Qwen3-32B2026.01 | 61.43 | |
| No ExperienceBase Model=Qwen3-8B2026.01 | 60 | |
| Training-Free GRPOBase Model=Qwen3-32B2026.01 | 59.57 | |
| REASONINGBANK+Base Model=Qwen3-8B2026.01 | 58.19 |