Deep Research on GAIA text-only original (test)
74.1Pass@1WebSailor-v2-30B-A3B (RL)
Evaluation Results
| Method | Links | |
|---|---|---|
| WebSailor-v2-30B-A3B (RL)Backbone Group=Medium Size Deep Research Agents (30B-72B), Training Protocol=RL2026.01 | 74.1 | |
| Tongyi-DeepResearchBackbone Group=Medium Size Deep Research Agents (30B-72B)2026.01 | 70.9 | |
| OpenAI-o3Backbone Group=Large Size Foundation Models (> 100B)2026.01 | 70.5 | |
| Claude-4-SonnetBackbone Group=Large Size Foundation Models (> 100B)2026.01 | 68.3 | |
| WebSailor-v2-30B-A3B (SFT)Backbone Group=Medium Size Deep Research Agents (30B-72B), Training Protocol=SFT2026.01 | 66 | |
| DeepSeek-V3.2Backbone Group=Large Size Foundation Models (> 100B)2026.01 | 63.5 | |
| DeepSeek-V3.1Backbone Group=Large Size Foundation Models (> 100B)2026.01 | 63.1 | |
| WebShaper-72BBackbone Group=Medium Size Deep Research Agents (30B-72B)2026.01 | 60 | |
| Kimi-K2Backbone Group=Large Size Foundation Models (> 100B)2026.01 | 57.7 | |
| MiroThinker-32B-DPO-v0.1Backbone Group=Medium Size Deep Research Agents (30B-72B), Training Protocol=DPO2026.01 | 57.3 | |
| WebSailor-72BBackbone Group=Medium Size Deep Research Agents (30B-72B)2026.01 | 55.4 | |
| WebSailor-32BBackbone Group=Medium Size Deep Research Agents (30B-72B)2026.01 | 53.2 | |
| ASearcher-Web-QwQBackbone Group=Medium Size Deep Research Agents (30B-72B)2026.01 | 52.8 | |
| WebShaper-32BBackbone Group=Medium Size Deep Research Agents (30B-72B)2026.01 | 52.4 | |
| WebDancer-QwQBackbone Group=Medium Size Deep Research Agents (30B-72B)2026.01 | 51.5 | |
| OffSeeker-8B (DPO)Backbone Group=Small Size Deep Research Agents (< 10B), Training Protocol=DPO2026.01 | 51.5 | |
| WebExplorer-8B (RL)Backbone Group=Small Size Deep Research Agents (< 10B), Training Protocol=RL2026.01 | 47.6 | |
| OffSeeker-8B (SFT)Backbone Group=Small Size Deep Research Agents (< 10B), Training Protocol=SFT2026.01 | 47.6 | |
| MiroThinker-8B-DPO-v0.1Backbone Group=Small Size Deep Research Agents (< 10B), Training Protocol=DPO2026.01 | 46.6 | |
| WebSailor-7BBackbone Group=Small Size Deep Research Agents (< 10B)2026.01 | 33 |