Multimodal Understanding on CCB (Chinese Comprehensive Benchmark)
75.3CCB ScoreHoVLE
Evaluation Results
| Method | Links | |
|---|---|---|
| HoVLEarchitecture=Monolithic, params=2.6B2024.12 | 75.3 | |
| InternVL2architecture=Compositional, params=2.2B2024.12 | 74.7 | |
| HoVLEarchitecture=Monolithic, params=2.6B, resolution=HD2024.12 | 74.3 | |
| InternVL1.5architecture=Compositional, params=2.2B2024.12 | 63.5 | |
| MiniCPM-V-2architecture=Compositional, params=2.8B2024.12 | 45.3 | |
| MiniCPM-Varchitecture=Compositional, params=2.8B2024.12 | 41.4 | |
| DeepSeek-VL-1.3Barchitecture=Compositional, params=2.0B2024.12 | 37.6 | |
| PaliGemma-3Barchitecture=Compositional, params=2.9B2024.12 | 29.6 | |
| VILA1.5-3Barchitecture=Compositional, params=3.1B2024.12 | 24.1 | |
| EVE-7Barchitecture=Monolithic, params=7B, resolution=HD2024.12 | 16.3 | |
| EVE-7Barchitecture=Monolithic, params=7B2024.12 | 12.4 | |
| Chameleon-7Barchitecture=Monolithic, params=7B2024.12 | 3.5 |