High-Resolution Visual Perception on HRBench 8K (Metrics: Overall, FSP, FCP)
75Overall ScoreActiveScope
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| ActiveScopeModel=Qwen3VL-30B-A3B(MoE)2026.06 | 75 | 89.75 | 60.25 | |
| ZoomEyeModel=Qwen3VL-30B-A3B(MoE)2026.06 | 73.88 | 88.25 | 59.5 | |
| RegularModel=Qwen3VL-30B-A3B(MoE)2026.06 | 73 | 86.75 | 59.25 | |
| DeepEyesbackbone=Qwen2.5-VL-7B2026.02 | 72.6 | 86.8 | 58.5 | |
| ActiveScopeModel=InternVL2.5-26B2026.06 | 71.25 | 79.25 | 63.25 | |
| Qwen2.5-VL-7BHEE=true2026.07 | 71.1 | 88.5 | 53.8 | |
| CapImaginebackbone=Qwen2.5-VL-7B2026.02 | 70.7 | 84.8 | 56.5 | |
| CapImaginebackbone=Qwen2.5-VL-7B, rewriting=false2026.02 | 69.8 | 83.5 | 56 | |
| ZoomEyeModel=InternVL2.5-26B2026.06 | 69.38 | 75.75 | 63 | |
| CapImaginebackbone=Qwen2.5-VL-7B, filtering=false2026.02 | 69.3 | 81.8 | 56.8 | |
| Monetbackbone=Qwen2.5-VL-7B, reported_from_prior_work=true2026.02 | 68 | 79.8 | 56.3 | |
| Monetbackbone=Qwen2.5-VL-7B, subset=true2026.02 | 67.9 | 84 | 51.8 | |
| LLaVA-ov-7BHEE=true2026.07 | 67.9 | 83.5 | 52.3 | |
| RegularModel=InternVL2.5-26B2026.06 | 67.38 | 73.25 | 61.5 | |
| PixelReasonerbackbone=Qwen2.5-VL-7B2026.02 | 66.9 | 80 | 54.3 | |
| InternVL2.5-8BHEE=true2026.07 | 66.5 | 81.8 | 51.3 | |
| Qwen2.5VL-7Bbackbone=Qwen2.5VL-7B2026.02 | 63.8 | 73.8 | 53.8 | |
| LVRbackbone=Qwen2.5-VL-7B2026.02 | 63 | 74.5 | 51.5 | |
| Qwen2.5-VL-7BHEE=false2026.07 | 62.1 | 73.3 | 51 | |
| InternVL3-8Bbackbone=InternVL3-8B2026.02 | 62 | 64.3 | 59.8 | |
| LLaVA-ov-7BHEE=false2026.07 | 57.3 | 66.5 | 48 | |
| GPT-4oreported_from_prior_work=true2026.02 | 55.5 | 62 | 49 | |
| GPT-4oHEE=false2026.07 | 55.5 | 62 | 49 | |
| InternVL2.5-8BHEE=false2026.07 | 54.8 | 59.5 | 50 | |
| Qwen-VL-MaxHEE=false2026.07 | 52.5 | 54 | 51 | |
| LLaVA-HR-X-13BHEE=false2026.07 | 46.9 | 49.5 | 44.3 | |
| LLaVA-HR-X-7BHEE=false2026.07 | 41.6 | 42 | 41.3 | |
| LLaVA-v1.6-7BHEE=false2026.07 | 40.8 | 37.3 | 44.3 | |
| Yi-VL-34BHEE=false2026.07 | 39 | 39.5 | 38.5 |