Long Video Understanding on L-VidBench (val)
66.7ScoreGPT-4o
Evaluation results
| Method | Links | |
|---|---|---|
| GPT-4o2026.03 | 66.7 | |
| VideoChat-FlashMax #F=10000, Max Res.=4482026.03 | 64.7 | |
| Gemini 1.5-Pro2026.03 | 64 | |
| InternVL3.5-8BMax #F=64, Max Res.=4482026.03 | 62.1 | |
| NVILA-8B-Video + AutoGazeMax #F=1024, Max Res.=35842026.03 | 61 | |
| VideoLLaMA3-7BMax #F=180, Max Res.=3842026.03 | 59.8 | |
| Apollo-7BMax #F=2FPS, Max Res.=3842026.03 | 58.5 | |
| LongVILA-R1-7BMax #F=8192, Max Res.=4482026.03 | 58 | |
| NVILA-8B-VideoMax #F=256, Max Res.=4482026.03 | 57.7 | |
| LongVILA-7BMax #F=2048, Max Res.=3842026.03 | 57.1 | |
| LLaVA-OV-8BMax #F=32, Max Res.=3842026.03 | 56.5 | |
| Qwen2.5-VL-7BMax #F=48, Max Res.=8962026.03 | 56 |