Multimodal Report Generation on WildSeek Text-Centric Complex Queries
53.89OrganizationCogGen (Ours)
Evaluation Results
| Method | Links | ||||||
|---|---|---|---|---|---|---|---|
| CogGen (Ours)Backbone=GPT-4.1, Search LLM=GPT-4.1-Mini, Retrieval Tool=Tavily Search, Temperature=0.52026.04 | 53.89 | 50 | 53.34 | 55.44 | 54.37 | 53.41 | |
| WriteHereBackbone=GPT-4.1, Retrieval Tool=Tavily Search2026.04 | 52.43 | 49.31 | 52.71 | 47.38 | 44.97 | 49.36 | |
| Gemini Deep Research (Ref)2026.04 | 50 | 50 | 50 | 50 | 50 | 50 | |
| STORMBackbone=GPT-4.1, Retrieval Tool=Tavily Search2026.04 | 43.75 | 40.97 | 44.72 | 19.03 | 19.08 | 33.51 | |
| Co-STORMBackbone=GPT-4.1, Retrieval Tool=Tavily Search2026.04 | 39.93 | 36.95 | 42.7 | 19.43 | 18.34 | 31.47 | |
| Multimodal DeepResearcherBackbone=GPT-4.1, Retrieval Tool=Tavily Search2026.04 | 38.19 | 37.4 | 36.95 | 20.76 | 21.83 | 31.03 |