Perception and Reasoning on RealWorldQA
74.2ScoreQwen3.5-4B
Evaluation Results
| Method | Links | |
|---|---|---|
| Qwen3.5-4BModel=Qwen3.5-4B2026.05 | 74.2 | |
| Molmo2-4BModel=Molmo2-4B2026.05 | 73.8 | |
| PLM-3BModel=PLM-3B2026.05 | 73.2 | |
| InternVL3.5-20B-A4BModel=InternVL3.5-20B-A4B2026.05 | 71.2 | |
| Qwen3-VL-4BModel=Qwen3-VL-4B2026.05 | 71 | |
| Qwen3-VL + LOCUSSize=4B2026.06 | 71 | |
| DeepSeek-VL2-16B-A2.4BModel=DeepSeek-VL2-16B-A2.4B2026.05 | 70 | |
| Qwen3-VLSize=4B2026.06 | 69.5 | |
| PixelReasonerVisual Bandwidth Constraint (B)=No Constrain, Tool-calling=True2026.05 | 69.02 | |
| Qwen3.5-2BModel=Qwen3.5-2B2026.05 | 69 | |
| DeepEyesVisual Bandwidth Constraint (B)=No Constrain, Tool-calling=True2026.05 | 68.89 | |
| Qwen2.5-VLVisual Bandwidth Constraint (B)=No Constrain, Reported From=Other publications2026.05 | 68.5 | |
| InternVL3.5-4BModel=InternVL3.5-4B2026.05 | 67.8 | |
| Starve to PerceiveVisual Bandwidth Constraint (B)=No Constrain, Tool-calling=True2026.05 | 67.71 | |
| Qwen2.5-VL + BA-SFTVisual Bandwidth Constraint (B)=No Constrain, Tool-calling=True2026.05 | 67.32 | |
| MiMo-VL + LOCUSSize=7B2026.06 | 67.3 | |
| GPT-4oVisual Bandwidth Constraint (B)=No Constrain, Reported From=Other publications2026.05 | 67.2 | |
| Qwen2.5-VL + BA-SFT + VanillaRLVisual Bandwidth Constraint (B)=No Constrain, Tool-calling=True2026.05 | 67.19 | |
| MiMo-VLSize=7B2026.06 | 66.7 | |
| Qwen2.5-VL + LOCUSSize=7B2026.06 | 66.4 | |
| LLaVA-OneVisionVisual Bandwidth Constraint (B)=No Constrain, Reported From=Other publications2026.05 | 66.3 | |
| Qwen3-VL-2BModel=Qwen3-VL-2B2026.05 | 66 | |
| Qwen2.5-VL-3BModel=Qwen2.5-VL-3B2026.05 | 65.6 | |
| ZAYA1-VL-8B-A1BModel=ZAYA1-VL-8B-A1B2026.05 | 65 | |
| Starve to PerceiveVisual Bandwidth Constraint (B)=256, Tool-calling=True2026.05 | 64.97 | |
| ChainOfFocusVisual Bandwidth Constraint (B)=No Constrain, Tool-calling=True2026.05 | 64.45 | |
| Qwen2.5-VL + BA-SFT + VanillaRLVisual Bandwidth Constraint (B)=256, Tool-calling=True2026.05 | 63.27 | |
| DeepEyesVisual Bandwidth Constraint (B)=256, Tool-calling=True2026.05 | 63.01 | |
| InternVL3.5-2BModel=InternVL3.5-2B2026.05 | 61.6 | |
| Qwen2.5-VLSize=7B2026.06 | 61.6 | |
| Qwen2.5-VL + BA-SFTVisual Bandwidth Constraint (B)=256, Tool-calling=True2026.05 | 60.65 | |
| MolmoE-8B-A1BModel=MolmoE-8B-A1B2026.05 | 60.4 | |
| ChainOfFocusVisual Bandwidth Constraint (B)=256, Tool-calling=True2026.05 | 60 | |
| Qwen2.5-VLVisual Bandwidth Constraint (B)=2562026.05 | 56.34 | |
| PixelReasonerVisual Bandwidth Constraint (B)=256, Tool-calling=True2026.05 | 55.03 | |
| Mini-o3Visual Bandwidth Constraint (B)=No Constrain, Tool-calling=True2026.05 | 52.55 | |
| Mini-o3Visual Bandwidth Constraint (B)=256, Tool-calling=True2026.05 | 48.63 |