Video Captioning on MCTS-VCB
84.9Appearance Description PrecisionGPT-4o
Evaluation results
| Method | Links | ||||||||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| GPT-4oModel Type=Closed-Source2025.06 | 84.9 | 47.2 | 60.6 | 86.4 | 62.7 | 72.7 | 80.9 | 67.1 | 73.4 | 87.3 | 50.6 | 64.1 | 81.9 | 40.8 | 54.5 | 83.8 | 57.5 | 68.2 | |
| LLaVa_Video_Qwen2_72BModel Type=Open-Source2025.06 | 79.8 | 45.8 | 58.2 | 80.9 | 57.9 | 67.5 | 85.2 | 59.1 | 69.8 | 80.2 | 41.5 | 54.7 | 77.9 | 44.1 | 56.3 | 82.1 | 51.7 | 63.4 | |
| Tarsier-34BModel Type=Open-Source2025.06 | 79.8 | 34.3 | 48 | 75.5 | 51.5 | 61.2 | 86.8 | 55.9 | 68 | 76.8 | 36.4 | 49.4 | 62.1 | 35.8 | 45.5 | 79.6 | 45.7 | 58 | |
| Gemini-1.5-ProModel Type=Closed-Source2025.06 | 78.2 | 59.9 | 67.9 | 82.4 | 64.8 | 72.5 | 81 | 70.7 | 75.5 | 82 | 56 | 66.5 | 79.9 | 59 | 67.9 | 80.9 | 63.6 | 71.2 | |
| LLaVa_Video_Qwen2_7BModel Type=Open-Source2025.06 | 78 | 41.3 | 54 | 79.7 | 50.7 | 61.9 | 82.7 | 58.4 | 68.5 | 76.4 | 41 | 53.4 | 75.6 | 53.9 | 62.9 | 79.7 | 48.9 | 60.6 | |
| Tarsier-7BModel Type=Open-Source2025.06 | 76.3 | 34 | 47.1 | 69.7 | 49.7 | 58 | 85.7 | 56.4 | 68 | 72.9 | 35.5 | 47.8 | 58.7 | 36 | 44.6 | 76.7 | 45.2 | 56.9 | |
| LLaVa_OV_Qwen2_7BModel Type=Open-Source2025.06 | 75.8 | 45.1 | 56.5 | 77.3 | 58.9 | 66.8 | 84.5 | 59.1 | 69.6 | 77.1 | 42.2 | 54.5 | 73.2 | 42.4 | 53.7 | 79.6 | 45.3 | 57.7 | |
| LLaVa_OV_Qwen2_72BModel Type=Open-Source2025.06 | 75.8 | 45.5 | 56.9 | 78.9 | 60.3 | 68.3 | 86.1 | 60.2 | 70.9 | 78.7 | 43.1 | 55.7 | 76.1 | 40.8 | 53.1 | 81.3 | 52.9 | 64.1 | |
| MiniCPM-V-2_6Model Type=Open-Source2025.06 | 75.3 | 40.8 | 52.9 | 73 | 57.8 | 64.5 | 82.7 | 58.2 | 68.3 | 78.3 | 38.9 | 52 | 70.1 | 40.9 | 51.7 | 77.7 | 50.4 | 61.2 | |
| VILA1.5-40BModel Type=Open-Source2025.06 | 74.5 | 42.3 | 54 | 76.1 | 55.8 | 64.4 | 82.1 | 53.4 | 64.7 | 79 | 37.9 | 51.2 | 73.1 | 39.8 | 51.6 | 78.3 | 47.6 | 59.2 | |
| Qwen2-VL-7BModel Type=Open-Source2025.06 | 74.4 | 43.5 | 54.9 | 75.8 | 57.7 | 65.5 | 80.4 | 60.2 | 68.8 | 77.6 | 40.7 | 53.4 | 71.8 | 34.9 | 46.9 | 77.8 | 50.9 | 61.5 | |
| PLLAVA-34BModel Type=Open-Source2025.06 | 72.3 | 33.5 | 45.8 | 73.2 | 44.6 | 55.4 | 78.3 | 46.6 | 58.4 | 73.6 | 27.9 | 40.4 | 65.6 | 26.5 | 37.7 | 74.9 | 38.5 | 50.9 | |
| Llama-3-VILA1.5-8BModel Type=Open-Source2025.06 | 71.7 | 39.2 | 50.7 | 73.8 | 46.6 | 57.1 | 80.9 | 52.3 | 63.5 | 73.5 | 37.1 | 49.3 | 66.9 | 31.6 | 42.9 | 75.6 | 44.2 | 55.8 | |
| Qwen2-VL-72BModel Type=Open-Source2025.06 | 71.4 | 45.8 | 55.8 | 75.9 | 57.9 | 65.7 | 80.9 | 60.8 | 69.4 | 78.7 | 42.5 | 55.2 | 73.2 | 37.7 | 49.7 | 77.7 | 52.1 | 62.4 | |
| PLLaVA-13BModel Type=Open-Source2025.06 | 69.9 | 35.9 | 47.4 | 69.5 | 44 | 53.9 | 78 | 46.2 | 58 | 66.7 | 27.3 | 38.8 | 69.8 | 29.5 | 41.4 | 72.5 | 38.7 | 50.4 | |
| VILA1.5-13BModel Type=Open-Source2025.06 | 69.4 | 39 | 50 | 70.5 | 50.3 | 58.7 | 79.3 | 51.6 | 62.5 | 74 | 34.6 | 47.2 | 66.8 | 35.2 | 46.1 | 73.9 | 44.4 | 55.4 | |
| InternVL2_5-78BModel Type=Open-Source2025.06 | 65.3 | 38.2 | 48.2 | 67.7 | 44.7 | 53.8 | 76.4 | 49.8 | 60.3 | 73.5 | 31.8 | 44.4 | 55.6 | 32 | 40.6 | 70.6 | 41.6 | 52.4 | |
| InternVL2_5-26BModel Type=Open-Source2025.06 | 65.2 | 37.1 | 47.3 | 67.6 | 43.4 | 52.8 | 75.6 | 49.1 | 59.5 | 73.6 | 31.5 | 44.1 | 54.4 | 32 | 40.3 | 70.6 | 41 | 51.8 | |
| InternVL2_5-8BModel Type=Open-Source2025.06 | 65.1 | 36.5 | 46.8 | 65 | 42 | 51 | 76 | 48.7 | 59.4 | 72.1 | 30.3 | 42.7 | 54.7 | 31.7 | 40.1 | 69.6 | 40.1 | 50.8 | |
| InternVL2_5-38BModel Type=Open-Source2025.06 | 65 | 37 | 47.1 | 68.2 | 44 | 53.5 | 75.7 | 48.9 | 59.5 | 73.8 | 31.7 | 44.4 | 56.2 | 31.7 | 40.6 | 70.6 | 41 | 51.8 | |
| PLLAVA-7BModel Type=Open-Source2025.06 | 47.4 | 28.3 | 35.5 | 47.6 | 37.6 | 42 | 69 | 37.7 | 48.7 | 61.9 | 17.6 | 27.4 | 37.6 | 20.3 | 26.4 | 55.1 | 30.4 | 39.2 |