3D Reasoning over Long Videos on HourVideo (dev)
39.2Overall AccuracyAria
Evaluation Results
| Method | Links | |
|---|---|---|
| Aria2025.05 | 39.2 | |
| Gemini 1.5 Pro2025.05 | 37.4 | |
| SiLVR2025.05 | 36.3 | |
| Qwen2-VL 7B2025.05 | 33.8 | |
| VAMBA2025.05 | 33.6 |
| Method | Links | |
|---|---|---|
| Aria2025.05 | 39.2 | |
| Gemini 1.5 Pro2025.05 | 37.4 | |
| SiLVR2025.05 | 36.3 | |
| Qwen2-VL 7B2025.05 | 33.8 | |
| VAMBA2025.05 | 33.6 |