Video Deep Research on VideoDR
77ScoreGemini-3-Pro
Evaluation Results
| Method | Links | |
|---|---|---|
| Gemini-3-ProWorkflow=Agentic Model, Zero-shot=true2026.07 | 77 | |
| Gemini-3-ProWorkflow=Direct Answer, Zero-shot=false2026.07 | 61 | |
| Gemini-3-FlashWorkflow=Agentic Model, Zero-shot=true2026.07 | 60 | |
| GPT-5.2Workflow=Agentic Model, Zero-shot=true2026.07 | 58 | |
| Gemini-3-FlashWorkflow=Direct Answer, Zero-shot=false2026.07 | 57 | |
| VideoSearcher-8BWorkflow=Agentic Model, Zero-shot=false2026.07 | 53 | |
| GPT-4oWorkflow=Agentic Model, Zero-shot=true2026.07 | 48 | |
| VideoSearcher-4BWorkflow=Agentic Model, Zero-shot=false2026.07 | 46 | |
| GPT-5.2Workflow=Direct Answer, Zero-shot=false2026.07 | 40 | |
| Webwatcher-7BWorkflow=Agentic Model, Zero-shot=false2026.07 | 37 | |
| Qwen3-VL-32B-InstructWorkflow=Agentic Model, Zero-shot=true2026.07 | 36 | |
| GPT-4oWorkflow=Direct Answer, Zero-shot=false2026.07 | 35 | |
| Qwen2.5-VL-32B-InstructWorkflow=Agentic Model, Zero-shot=true2026.07 | 30 | |
| Qwen3-VL-8B-InstructWorkflow=Agentic Model, Zero-shot=true2026.07 | 28 | |
| Qwen3-VL-4B-InstructWorkflow=Agentic Model, Zero-shot=true2026.07 | 25 | |
| DeepEyesV2Workflow=Agentic Model, Zero-shot=false2026.07 | 19 | |
| Qwen2.5-VL-32B-InstructWorkflow=Direct Answer, Zero-shot=false2026.07 | 16 | |
| Qwen3-VL-32B-InstructWorkflow=Direct Answer, Zero-shot=false2026.07 | 16 | |
| Qwen3-VL-4B-InstructWorkflow=Direct Answer, Zero-shot=false2026.07 | 8 | |
| Qwen3-VL-8B-InstructWorkflow=Direct Answer, Zero-shot=false2026.07 | 8 | |
| Qwen2.5-VL-7B-InstructWorkflow=Agentic Model, Zero-shot=true2026.07 | 7 | |
| Qwen2.5-VL-7B-InstructWorkflow=Direct Answer, Zero-shot=false2026.07 | 5 | |
| Visual-ARFTWorkflow=Agentic Model, Zero-shot=false2026.07 | 5 | |
| MMSearch-R1Workflow=Agentic Model, Zero-shot=false2026.07 | 5 |