Video Question Answering on LONGVIDEOBENCH Medium
56.1AccuracyGPT-4O + VSI
Evaluation Results
| Method | Links | |
|---|---|---|
| GPT-4O + VSISearching Modality=Multimodal, Frame=82025.08 | 56.1 | |
| GPT-4O + VSISearching Modality=Multimodal, Frame=322025.08 | 53.9 | |
| GPT-4O + TstarSearching Modality=Unimodal, Frame=82025.08 | 52.9 | |
| GPT-4O + TstarSearching Modality=Unimodal, Frame=322025.08 | 52.5 | |
| GPT-4O + VSLSSearching Modality=Unimodal, Frame=82025.08 | 52.2 | |
| GPT-4OSearching Modality=N/A, Frame=322025.08 | 51.9 | |
| GPT-4O + VSLSSearching Modality=Unimodal, Frame=322025.08 | 50.7 | |
| GPT-4OSearching Modality=N/A, Frame=82025.08 | 50.5 | |
| LLaVA-Video-7B-Qwen2Searching Modality=N/A, Frame=322025.08 | 45.3 | |
| LLaVA-Video-7B-Qwen2+VSISearching Modality=Multimodal, Frame=322025.08 | 45.3 | |
| QWEN2.5-VL-7B-INSTRUCT+VSISearching Modality=Multimodal, Frame=82025.08 | 45.1 | |
| QWEN2.5-VL-7B-INSTRUCT+VSISearching Modality=Multimodal, Frame=322025.08 | 44.4 | |
| LLaVA-Video-7B-Qwen2+TstarSearching Modality=Unimodal, Frame=82025.08 | 44.1 | |
| LLaVA-Video-7B-Qwen2+VSISearching Modality=Multimodal, Frame=82025.08 | 44.1 | |
| QWEN2.5-VL-7B-INSTRUCT+TstarSearching Modality=Unimodal, Frame=82025.08 | 43.9 | |
| QWEN2.5-VL-7B-INSTRUCT+VSLSSearching Modality=Unimodal, Frame=82025.08 | 43.9 | |
| LLaVA-Video-7B-Qwen2+TstarSearching Modality=Unimodal, Frame=322025.08 | 43.9 | |
| LLaVA-Video-7B-Qwen2+VSLSSearching Modality=Unimodal, Frame=322025.08 | 43.9 | |
| QWEN2.5-VL-7B-INSTRUCTSearching Modality=N/A, Frame=82025.08 | 43.5 | |
| QWEN2.5-VL-7B-INSTRUCT+VSLSSearching Modality=Unimodal, Frame=322025.08 | 43.1 | |
| QWEN2.5-VL-7B-INSTRUCT+TstarSearching Modality=Unimodal, Frame=322025.08 | 42.9 | |
| LLaVA-Video-7B-Qwen2Searching Modality=N/A, Frame=82025.08 | 42.7 | |
| LLaVA-Video-7B-Qwen2+VSLSSearching Modality=Unimodal, Frame=82025.08 | 42.4 | |
| QWEN2.5-VL-7B-INSTRUCTSearching Modality=N/A, Frame=322025.08 | 35 |