Visual Search and Perception-intensive Reasoning on HR-Bench 4K
76.63Overall ScoreStarve to Perceive
Evaluation Results
| Method | Links | |
|---|---|---|
| Starve to PerceiveVisual Bandwidth Constraint (B)=No Constrain, Tool-calling=True2026.05 | 76.63 | |
| Qwen2.5-VL + BA-SFT + VanillaRLVisual Bandwidth Constraint (B)=No Constrain, Tool-calling=True2026.05 | 75.25 | |
| Qwen2.5-VL + BA-SFTVisual Bandwidth Constraint (B)=No Constrain, Tool-calling=True2026.05 | 72.5 | |
| ChainOfFocusVisual Bandwidth Constraint (B)=No Constrain, Tool-calling=True2026.05 | 71.63 | |
| DeepEyesVisual Bandwidth Constraint (B)=No Constrain, Tool-calling=True2026.05 | 71.5 | |
| PixelReasonerVisual Bandwidth Constraint (B)=No Constrain, Tool-calling=True2026.05 | 71.13 | |
| GPT-4oVisual Bandwidth Constraint (B)=No Constrain, Reported From=Other publications2026.05 | 68 | |
| Qwen2.5-VL + BA-SFT + VanillaRLVisual Bandwidth Constraint (B)=256, Tool-calling=True2026.05 | 66.25 | |
| Starve to PerceiveVisual Bandwidth Constraint (B)=256, Tool-calling=True2026.05 | 66.13 | |
| Qwen2.5-VLVisual Bandwidth Constraint (B)=No Constrain, Reported From=Other publications2026.05 | 65.25 | |
| LLaVA-OneVisionVisual Bandwidth Constraint (B)=No Constrain, Reported From=Other publications2026.05 | 63 | |
| ChainOfFocusVisual Bandwidth Constraint (B)=256, Tool-calling=True2026.05 | 59.63 | |
| Qwen2.5-VL + BA-SFTVisual Bandwidth Constraint (B)=256, Tool-calling=True2026.05 | 59.25 | |
| Mini-o3Visual Bandwidth Constraint (B)=No Constrain, Tool-calling=True2026.05 | 57.75 | |
| DeepEyesVisual Bandwidth Constraint (B)=256, Tool-calling=True2026.05 | 57.13 | |
| Mini-o3Visual Bandwidth Constraint (B)=256, Tool-calling=True2026.05 | 51.63 | |
| PixelReasonerVisual Bandwidth Constraint (B)=256, Tool-calling=True2026.05 | 49.38 | |
| Qwen2.5-VLVisual Bandwidth Constraint (B)=2562026.05 | 38 |