Question Answering on LongVideoBench Long Text-Related Perception
69.57AccuracyGPT4o+VSI
Evaluation Results
| Method | Links | |
|---|---|---|
| GPT4o+VSIFrame=82025.08 | 69.57 | |
| GPT4o+SUBFrame=82025.08 | 58.06 | |
| GPT4o+TSTARFrame=82025.08 | 58.06 | |
| GPT4o+VSLSFrame=82025.08 | 55.91 | |
| GPT4oFrame=82025.08 | 53.76 |