Visual Question Answering on HRBench-4K (FSP/FCP/Overall)
92.8FSP ScoreConFoThinking (Qwen3-VL-8B)
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| ConFoThinking (Qwen3-VL-8B)Backbone=Qwen3-VL-8B2026.02 | 92.8 | 67.3 | 80 | |
| ConFoThinking (Qwen2.5-VL-7B)Backbone=Qwen2.5-VL-7B2026.02 | 91.8 | 56.8 | 74.3 | |
| Qwen3-VL-8BBackbone=Qwen3-VL-8B2026.02 | 88.3 | 64.5 | 76.4 | |
| ConFoThinking (Qwen3-VL-4B)Backbone=Qwen3-VL-4B2026.02 | 87.8 | 68 | 77.9 | |
| Pixel-Reasoner2026.02 | 86.3 | 61.8 | 74 | |
| Qwen3-VL-4BBackbone=Qwen3-VL-4B2026.02 | 86 | 62.5 | 74.3 | |
| LLaVA-OneVision-1.5-8BBackbone=LLaVA-OneVision-1.5-8B2026.02 | 85.8 | 58.2 | 72 | |
| Qwen2.5-VL-7BBackbone=Qwen2.5-VL-7B2026.02 | 85.3 | 52.3 | 68.8 | |
| InternVL3.5-8BBackbone=InternVL3.5-8B2026.02 | 85.3 | 60.8 | 73.2 | |
| Qwen3-VL-8B (with tools)Backbone=Qwen3-VL-8B, Thinking Tools=true2026.02 | 85 | 65.5 | 75.3 | |
| Qwen3-VL-4B (with tools)Backbone=Qwen3-VL-4B, Thinking Tools=true2026.02 | 84.3 | 62.3 | 73.3 | |
| ZoomEye2026.02 | 84.2 | 55 | 69.6 | |
| Visual Sketchpad2026.02 | 81.3 | 57.8 | 69.5 | |
| SEAL2026.02 | 71.5 | 49 | 60.3 | |
| ICoT2026.02 | 65.5 | 55 | 60.3 |