College-level Multimodal Understanding on MMMU
41.1MMMU ScoreQwen2-VL-2B
Evaluation Results
| Method | Links | |
|---|---|---|
| Qwen2-VL-2Barchitecture=Compositional, params=2.2B2024.12 | 41.1 | |
| Phi-3 Visionarchitecture=Compositional, params=4.2B2024.12 | 40.4 | |
| MiniCPM-Varchitecture=Compositional, params=2.8B2024.12 | 38.3 | |
| MiniCPM-V-2architecture=Compositional, params=2.8B2024.12 | 38.2 | |
| PaliGemma-3Barchitecture=Compositional, params=2.9B2024.12 | 34.9 | |
| InternVL1.5architecture=Compositional, params=2.2B2024.12 | 34.6 | |
| InternVL2architecture=Compositional, params=2.2B2024.12 | 34.3 | |
| MM1-3B-Chatarchitecture=Compositional, params=3B2024.12 | 33.9 | |
| HoVLEarchitecture=Monolithic, params=2.6B2024.12 | 33.7 | |
| VILA1.5-3Barchitecture=Compositional, params=3.1B2024.12 | 33.3 | |
| EVE-7Barchitecture=Monolithic, params=7B, resolution=HD2024.12 | 32.6 | |
| EVE-7Barchitecture=Monolithic, params=7B2024.12 | 32.3 | |
| DeepSeek-VL-1.3Barchitecture=Compositional, params=2.0B2024.12 | 32.2 | |
| HoVLEarchitecture=Monolithic, params=2.6B, resolution=HD2024.12 | 32.2 | |
| Emu3architecture=Monolithic, params=8B2024.12 | 31.6 | |
| Chameleon-7Barchitecture=Monolithic, params=7B2024.12 | 25.4 |