Visual Grounding on HRBench8K
75.63AccuracyERASE
Evaluation Results
| Method | Links | ||||
|---|---|---|---|---|---|
| ERASERatio=-65.11%, Backbone=Qwen3-VL-8B2026.05 | 75.63 | — | — | — | |
| ERASERatio=-84.43%, Backbone=Qwen3-VL-8B2026.05 | 74.88 | — | — | — | |
| ERASERatio=-72.77%, Backbone=Qwen3-VL-8B2026.05 | 74.75 | — | — | — | |
| IVC-PruneRatio=-65.11%, Backbone=Qwen3-VL-8B2026.05 | 72.75 | — | — | — | |
| PruneSIDRatio=-72.77%, Backbone=Qwen3-VL-8B2026.05 | 72.38 | — | — | — | |
| BaseRatio=100%, Backbone=Qwen3-VL-8B2026.05 | 72.13 | — | — | — | |
| IVC-PruneRatio=-72.77%, Backbone=Qwen3-VL-8B2026.05 | 72.13 | — | — | — | |
| IVC-PruneRatio=-84.43%, Backbone=Qwen3-VL-8B2026.05 | 72.13 | — | — | — | |
| PruneSIDRatio=-84.43%, Backbone=Qwen3-VL-8B2026.05 | 71.63 | — | — | — | |
| DARTRatio=-65.11%, Backbone=Qwen3-VL-8B2026.05 | 71.12 | — | — | — | |
| PruneSIDRatio=-65.11%, Backbone=Qwen3-VL-8B2026.05 | 71.12 | — | — | — | |
| iLLaVARatio=-65.11%, Backbone=Qwen3-VL-8B2026.05 | 70.88 | — | — | — | |
| iLLaVARatio=-72.77%, Backbone=Qwen3-VL-8B2026.05 | 70.25 | — | — | — | |
| DARTRatio=-62.5%, Backbone=InternVL3-8B2026.05 | 69.62 | — | — | — | |
| DARTRatio=-62.5%, Backbone=InternVL3-8B2026.05 | 69.62 | — | — | — | |
| BaseRatio=100%, Backbone=InternVL3-8B2026.05 | 68.75 | — | — | — | |
| ERASERatio=-75.0%, Backbone=InternVL3-8B2026.05 | 68.75 | — | — | — | |
| ERASERatio=-85.0%, Backbone=InternVL3-8B2026.05 | 68.75 | — | — | — | |
| BaseRatio=100%, Backbone=InternVL3-8B2026.05 | 68.75 | — | — | — | |
| ERASERatio=-75.0%, Backbone=InternVL3-8B2026.05 | 68.75 | — | — | — | |
| ERASERatio=-85.0%, Backbone=InternVL3-8B2026.05 | 68.75 | — | — | — | |
| IVC-PruneRatio=-62.5%, Backbone=InternVL3-8B2026.05 | 68.63 | — | — | — | |
| IVC-PruneRatio=-62.5%, Backbone=InternVL3-8B2026.05 | 68.63 | — | — | — | |
| DARTRatio=-72.77%, Backbone=Qwen3-VL-8B2026.05 | 68.5 | — | — | — | |
| IVC-PruneRatio=-75.0%, Backbone=InternVL3-8B2026.05 | 68.5 | — | — | — | |
| IVC-PruneRatio=-75.0%, Backbone=InternVL3-8B2026.05 | 68.5 | — | — | — | |
| iLLaVARatio=-84.43%, Backbone=Qwen3-VL-8B2026.05 | 68.25 | — | — | — | |
| ERASERatio=-62.5%, Backbone=InternVL3-8B2026.05 | 68.25 | — | — | — | |
| ERASERatio=-62.5%, Backbone=InternVL3-8B2026.05 | 68.25 | — | — | — | |
| DARTRatio=-75.0%, Backbone=InternVL3-8B2026.05 | 67.13 | — | — | — | |
| DARTRatio=-75.0%, Backbone=InternVL3-8B2026.05 | 67.13 | — | — | — | |
| PruneSIDRatio=-62.5%, Backbone=InternVL3-8B2026.05 | 66.75 | — | — | — | |
| PruneSIDRatio=-62.5%, Backbone=InternVL3-8B2026.05 | 66.75 | — | — | — | |
| IVC-PruneRatio=-85.0%, Backbone=InternVL3-8B2026.05 | 66.63 | — | — | — | |
| IVC-PruneRatio=-85.0%, Backbone=InternVL3-8B2026.05 | 66.63 | — | — | — | |
| CDPrunerRatio=-72.77%, Backbone=Qwen3-VL-8B2026.05 | 66.5 | — | — | — | |
| CDPrunerRatio=-65.11%, Backbone=Qwen3-VL-8B2026.05 | 65.38 | — | — | — | |
| DARTRatio=-85.0%, Backbone=InternVL3-8B2026.05 | 65 | — | — | — | |
| DARTRatio=-85.0%, Backbone=InternVL3-8B2026.05 | 65 | — | — | — | |
| CDPrunerRatio=-84.43%, Backbone=Qwen3-VL-8B2026.05 | 64.63 | — | — | — | |
| DARTRatio=-84.43%, Backbone=Qwen3-VL-8B2026.05 | 63.88 | — | — | — | |
| CDPrunerRatio=-62.5%, Backbone=InternVL3-8B2026.05 | 63.75 | — | — | — | |
| CDPrunerRatio=-62.5%, Backbone=InternVL3-8B2026.05 | 63.75 | — | — | — | |
| PruneSIDRatio=-75.0%, Backbone=InternVL3-8B2026.05 | 62.87 | — | — | — | |
| PruneSIDRatio=-75.0%, Backbone=InternVL3-8B2026.05 | 62.87 | — | — | — | |
| CDPrunerRatio=-75.0%, Backbone=InternVL3-8B2026.05 | 61.63 | — | — | — | |
| CDPrunerRatio=-75.0%, Backbone=InternVL3-8B2026.05 | 61.63 | — | — | — | |
| CDPrunerRatio=-85.0%, Backbone=InternVL3-8B2026.05 | 58.12 | — | — | — | |
| CDPrunerRatio=-85.0%, Backbone=InternVL3-8B2026.05 | 58.12 | — | — | — | |
| PruneSIDRatio=-85.0%, Backbone=InternVL3-8B2026.05 | 56.88 | — | — | — | |
| PruneSIDRatio=-85.0%, Backbone=InternVL3-8B2026.05 | 56.88 | — | — | — | |
| Base (InternVL3-8B-Instruct)Backbone=InternVL3-8B-Instruct, Training Free=false2026.06 | — | 80 | 59.8 | 69.9 | |
| Base (Qwen2.5-VL-7B-Instruct)Backbone=Qwen2.5-VL-7B-Instruct, Training Free=false2026.06 | — | 84.2 | 51.5 | 67.9 | |
| Base (Qwen3-VL-8B-Instruct)Backbone=Qwen3-VL-8B-Instruct, Training Free=false2026.06 | — | 83.8 | 65 | 74.4 | |
| DeepEyesBackbone=Qwen2.5-VL-7B-Instruct, Training Free=false2026.06 | — | 86.8 | 58.5 | 72.6 | |
| DeepScanBackbone=Qwen2.5-VL-7B-Instruct, Training Free=true2026.06 | — | 87.2 | 57.6 | 72.4 | |
| DyfoBackbone=Qwen2.5-VL-7B-Instruct, Training Free=true2026.06 | — | 86.5 | 53.2 | 69.8 | |
| GPT-4oTraining Free=true2026.06 | — | 62 | 49 | 55.5 | |
| HiDeBackbone=InternVL3-8B-Instruct, Training Free=true2026.06 | — | 86 | 54.2 | 70.1 | |
| HiDeBackbone=Qwen2.5-VL-7B-Instruct, Training Free=true2026.06 | — | 92.5 | 57.3 | 74.9 | |
| LazyMCoTBackbone=InternVL3-8B-Instruct, Training Free=true2026.06 | — | 89.8 | 53 | 71.4 | |
| LazyMCoTBackbone=Qwen3-VL-8B-Instruct, Training Free=true2026.06 | — | 90.8 | 64 | 77.4 | |
| LazyMCoTBackbone=Qwen2.5-VL-7B-Instruct, Training Free=true2026.06 | — | 93.8 | 57 | 75.4 | |
| Thyme-VLBackbone=Qwen2.5-VL-7B-Instruct, Training Free=false2026.06 | — | 86.5 | 57.5 | 72 | |
| TreeVGRBackbone=Qwen2.5-VL-7B-Instruct, Training Free=false2026.06 | — | 84.4 | 57.2 | 69.8 | |
| ViCropBackbone=InternVL3-8B-Instruct, Training Free=true2026.06 | — | 82.8 | 54.8 | 68.8 | |
| ViCropBackbone=Qwen2.5-VL-7B-Instruct, Training Free=true2026.06 | — | 85.5 | 53 | 69.3 | |
| ZoomRefineBackbone=Qwen2.5-VL-7B-Instruct, Training Free=true2026.06 | — | 83.9 | 54 | 68.6 |