Multimodal Understanding on AI2D w/o M
81.6AccuracyQwen2.5VL-3B
Evaluation Results
| Method | Links | |
|---|---|---|
| Qwen2.5VL-3BComplexity=O(n)2025.12 | 81.6 | |
| InternVL2.5-4BComplexity=O(n)2025.12 | 81.4 | |
| Phi-3.5-Vision-4BComplexity=O(n)2025.12 | 77.8 | |
| InfiniteVL-4BComplexity=O(1)2025.12 | 77.2 | |
| SmolVLM2-2BComplexity=O(n)2025.12 | 74.9 | |
| PaliGemma2-3BComplexity=O(n)2025.12 | 72.2 | |
| TinyLLaVA-3BComplexity=O(n)2025.12 | 61.8 | |
| MaTVLM-3BComplexity=O(n)2025.12 | 58.9 | |
| Cobra-3BComplexity=O(1)2025.12 | 46.8 |