3D/4D Video Question Answering on MMSI-Bench
40.7AccuracyGPT-5
Evaluation Results
| Method | Links | |
|---|---|---|
| GPT-52025.12 | 40.7 | |
| Gemini-2.5-Pro2025.12 | 36.9 | |
| 4D-RGPT-8Bparameters=8B2025.12 | 33.3 | |
| NVILA-Lite-8Bparameters=8B2025.12 | 31.3 | |
| GPT-4o2025.12 | 30.3 | |
| ViLaSR-7Bparameters=7B2025.12 | 30.2 | |
| SpaceR-7Bparameters=7B2025.12 | 28.8 | |
| InternVL2.5-8Bparameters=8B2025.12 | 28.7 | |
| LLaVA-NeXT-Video-7Bparameters=7B2025.12 | 26.8 | |
| Qwen2.5-VL-7Bparameters=7B2025.12 | 25.9 | |
| LLaVA-OneVision-7Bparameters=7B2025.12 | 24.5 | |
| SpatialReasoner-7Bparameters=7B2025.12 | 22.7 |