Web Search on xbench DeepSearch 2505
83.7ScoreStep 3.5 Flash
Evaluation results
| Method | Links | ||
|---|---|---|---|
| Step 3.5 FlashAgent Category=Large-scale Open-source Agents2026.06 | 83.7 | — | |
| FORT-SearcherAgent Category=Comparable-size Open-source Agents2026.06 | 80.8 | 66.2 | |
| OpenSeekerV2Agent Category=Comparable-size Open-source Agents2026.06 | 78 | 53.4 | |
| Qwen3.5-35B-A3BAgent Category=Comparable-size Open-source Agents2026.06 | 77.4 | 59.9 | |
| MiroThinker-1.7-miniAgent Category=Comparable-size Open-source Agents2026.06 | 77.2 | 64.6 | |
| Tongyi DeepResearchAgent Category=Comparable-size Open-source Agents2026.06 | 75 | 51.7 |