Fine-grained visual search on HR-Bench 8K
78Overall ScoreQwen3-VL + V-ABS
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| Qwen3-VL + V-ABSBackbone=Qwen3-VL, Plugin=V-ABS2026.05 | 78 | 88 | 68 | |
| Qwen2.5-VL + V-ABSBackbone=Qwen2.5-VL, Plugin=V-ABS2026.05 | 77.5 | 89 | 66 | |
| FOCUSexternal_tool=true2026.05 | 76.3 | — | — | |
| SeProDbase_method=Mini-o3, external_tool=false, prophet_model_size=7B2026.05 | 75.2 | 90.5 | 59.9 | |
| SeProDbase_method=Mini-o3, external_tool=false, prophet_model_size=3B2026.05 | 75 | 90.2 | 59.7 | |
| Qwen2.5-VL + DeepEyes v2Backbone=Qwen2.5-VL, Plugin=DeepEyes v22026.05 | 73.8 | — | — | |
| Mini-o3external_tool=false2026.05 | 73 | 88.9 | 57 | |
| Qwen2.5-VL + ThymeBackbone=Qwen2.5-VL, Plugin=Thyme2026.05 | 72 | 86.5 | 57.5 | |
| SeProDbase_method=DeepEyes, external_tool=false, prophet_model_size=3B2026.05 | 71.9 | 85.8 | 58 | |
| DeepEyesexternal_tool=false2026.05 | 69.9 | 85.3 | 54.5 | |
| ZoomEyesexternal_tool=true2026.05 | 69.3 | 88.5 | 50 | |
| Qwen3-VLModel=Qwen3-VL2026.05 | 68.5 | 82 | 55 | |
| Qwen2.5-VL + ZoomEyeBackbone=Qwen2.5-VL, Plugin=ZoomEye2026.05 | 68.4 | 84.8 | 52 | |
| Qwen2.5-VL + Pixel-ReasonerBackbone=Qwen2.5-VL, Plugin=Pixel-Reasoner2026.05 | 66.9 | — | — | |
| Intern-VL3 + V-ABSBackbone=Intern-VL3, Plugin=V-ABS2026.05 | 66 | 72 | 60 | |
| SeProDbase_method=Pixel Reasoner, external_tool=false, prophet_model_size=3B2026.05 | 65.1 | 77.5 | 52.8 | |
| GPT-4o + V-ABSBackbone=GPT-4o, Plugin=V-ABS2026.05 | 64.5 | 74.8 | 54.1 | |
| Pixel Reasonerexternal_tool=false2026.05 | 64.3 | 76.3 | 52.3 | |
| Qwen2.5-VL-7BModel=Qwen2.5-VL, Parameters=7B2026.05 | 62 | 81 | 43 | |
| Intern-VL3Model=Intern-VL32026.05 | 60.5 | 61 | 60 | |
| GPT-4oModel=GPT-4o2026.05 | 58 | 49 | 67 | |
| RAPexternal_tool=true2026.05 | 53.8 | 72.3 | 35.3 | |
| LLaVA-HR-X-7BModel=LLaVA-HR-X, Parameters=7B2026.05 | 41.6 | 42 | 41.3 | |
| LLaVA-v1.6-13BModel=LLaVA-v1.6, Parameters=13B2026.05 | 38.1 | 38 | 38.3 |