High-Resolution Multi-modal Search on HR-MMSearch
48.52ScoreGemini-3-Pro
Evaluation Results
| Method | Links | |
|---|---|---|
| Gemini-3-ProWorkflow=Agentic Model, Zero-shot=true2026.07 | 48.52 | |
| GPT-5.2Workflow=Agentic Model, Zero-shot=true2026.07 | 48.2 | |
| VideoSearcher-8BWorkflow=Agentic Model, Zero-shot=false2026.07 | 43.61 | |
| Gemini-3-FlashWorkflow=Agentic Model, Zero-shot=true2026.07 | 41.64 | |
| VideoSearcher-4BWorkflow=Agentic Model, Zero-shot=false2026.07 | 39.67 | |
| Qwen3-VL-32B-InstructWorkflow=Agentic Model, Zero-shot=true2026.07 | 34.43 | |
| Qwen2.5-VL-32B-InstructWorkflow=Agentic Model, Zero-shot=true2026.07 | 33.44 | |
| GPT-4oWorkflow=Agentic Model, Zero-shot=true2026.07 | 30.16 | |
| Qwen3-VL-8B-InstructWorkflow=Agentic Model, Zero-shot=true2026.07 | 27.87 | |
| Gemini-3-ProWorkflow=Direct Answer, Zero-shot=false2026.07 | 26.89 | |
| Webwatcher-7BWorkflow=Agentic Model, Zero-shot=false2026.07 | 26.89 | |
| GPT-5.2Workflow=Direct Answer, Zero-shot=false2026.07 | 24.92 | |
| Qwen3-VL-4B-InstructWorkflow=Agentic Model, Zero-shot=true2026.07 | 24.92 | |
| Visual-ARFTWorkflow=Agentic Model, Zero-shot=false2026.07 | 24.92 | |
| Gemini-3-FlashWorkflow=Direct Answer, Zero-shot=false2026.07 | 21.97 | |
| MMSearch-R1Workflow=Agentic Model, Zero-shot=false2026.07 | 20.33 | |
| Qwen2.5-VL-7B-InstructWorkflow=Agentic Model, Zero-shot=true2026.07 | 19.34 | |
| Qwen3-VL-32B-InstructWorkflow=Direct Answer, Zero-shot=false2026.07 | 19.02 | |
| DeepEyesV2Workflow=Agentic Model, Zero-shot=false2026.07 | 14.1 | |
| GPT-4oWorkflow=Direct Answer, Zero-shot=false2026.07 | 13.11 | |
| Qwen3-VL-8B-InstructWorkflow=Direct Answer, Zero-shot=false2026.07 | 12.13 | |
| Qwen2.5-VL-32B-InstructWorkflow=Direct Answer, Zero-shot=false2026.07 | 3.93 | |
| Qwen3-VL-4B-InstructWorkflow=Direct Answer, Zero-shot=false2026.07 | 3.61 | |
| Qwen2.5-VL-7B-InstructWorkflow=Direct Answer, Zero-shot=false2026.07 | 0.58 |