Multimodal Tasks on MMT-Bench 1.0 (test)
63.4OverallInternVL 1.2
Evaluation results
| Method | Links | ||
|---|---|---|---|
| InternVL 1.2open-source=true, param=40B2024.04 | 63.4 | 58.2 | |
| Qwen-VL-Plusopen-source=false2024.04 | 62.3 | 56.6 | |
| GPT-4Vopen-source=false2024.04 | 62 | 55.5 | |
| Gemini Pro 1.0open-source=false2024.04 | 61.6 | 55.1 | |
| LLaVA-NeXTopen-source=true, param=35B2024.04 | 60.8 | 56.3 | |
| InternVL 1.5open-source=true, param=26B2024.04 | 59 | 56.2 | |
| XComposer2open-source=true, param=8B2024.04 | 55.7 | 50 | |
| BLIP-2-XXLopen-source=true, param=12B2024.04 | 54.8 | 49.1 | |
| Yi-VL-34Bopen-source=true, param=35B2024.04 | 54.2 | 48.6 | |
| Monkey-Chatopen-source=true, param=10B2024.04 | 53.4 | 46 | |
| DeepSeek-VLopen-source=true, param=7B2024.04 | 53.2 | 46.5 | |
| Claude-3 Haikuopen-source=false2024.04 | 52.2 | 46.4 | |
| CogVLM-Chatopen-source=true, param=17B2024.04 | 51.6 | 44.2 |