Multi-step Reasoning on BrowseComp-Plus OOD
54.6Success Rate (SR)Qwen3-Emb + LRAT
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| Qwen3-Emb + LRATAgent Backbone=GLM-4.7 (358B)2026.03 | 54.6 | 77.8 | 44.6 | |
| E5-Large + LRATAgent Backbone=GLM-4.7 (358B)2026.03 | 50.6 | 76.3 | 44.8 | |
| E5-Large + LRATAgent Backbone=MiniMax-M2.1 (229B)2026.03 | 48.7 | 69.7 | 28.9 | |
| Qwen3-Emb + LRATAgent Backbone=MiniMax-M2.1 (229B)2026.03 | 48.3 | 69.2 | 28.3 | |
| E5-LargeAgent Backbone=MiniMax-M2.1 (229B)2026.03 | 46.4 | 64.9 | 29.1 | |
| E5-LargeAgent Backbone=GLM-4.7 (358B)2026.03 | 46.4 | 68.7 | 44.6 | |
| Qwen3-EmbAgent Backbone=GLM-4.7 (358B)2026.03 | 43.9 | 66.6 | 45.5 | |
| Qwen3-EmbAgent Backbone=MiniMax-M2.1 (229B)2026.03 | 38.2 | 57.2 | 30.8 | |
| E5-Large + LRATAgent Backbone=WebExplore (8B)2026.03 | 29 | 56.1 | 39.1 | |
| Qwen3-Emb + LRATAgent Backbone=WebExplore (8B)2026.03 | 27.2 | 55.9 | 38.7 | |
| E5-LargeAgent Backbone=WebExplore (8B)2026.03 | 25.4 | 50.4 | 40.1 | |
| E5-Large + LRATAgent Backbone=Tongyi-DeepResearch (30B)2026.03 | 23.9 | 61.8 | 41.4 | |
| Qwen3-Emb + LRATAgent Backbone=Tongyi-DeepResearch (30B)2026.03 | 23.7 | 60.7 | 41 | |
| Qwen3-EmbAgent Backbone=WebExplore (8B)2026.03 | 21 | 47.7 | 40.7 | |
| E5-LargeAgent Backbone=Tongyi-DeepResearch (30B)2026.03 | 20.7 | 54.8 | 42.4 | |
| Qwen3-EmbAgent Backbone=Tongyi-DeepResearch (30B)2026.03 | 17.8 | 49.2 | 42.9 | |
| E5-LargeAgent Backbone=AgentCPM-Explore (4B)2026.03 | 15.9 | 26.5 | 40.7 | |
| E5-Large + LRATAgent Backbone=AgentCPM-Explore (4B)2026.03 | 15.9 | 32.1 | 40.1 | |
| Qwen3-Emb + LRATAgent Backbone=AgentCPM-Explore (4B)2026.03 | 15.8 | 32 | 40.4 | |
| Qwen3-EmbAgent Backbone=AgentCPM-Explore (4B)2026.03 | 13.5 | 23.2 | 40.7 | |
| E5-Large + LRATAgent Backbone=GPT-OSS (120B)2026.03 | 13.1 | 56 | 44.6 | |
| Qwen3-Emb + LRATAgent Backbone=GPT-OSS (120B)2026.03 | 12.1 | 56.4 | 45.2 | |
| E5-LargeAgent Backbone=GPT-OSS (120B)2026.03 | 10.8 | 50.1 | 44.8 | |
| Qwen3-EmbAgent Backbone=GPT-OSS (120B)2026.03 | 9 | 43.7 | 45.4 |