3D Question Answering on ScanQA v1.0 (val)
16.2BLEU-4Efficient-VLN
Evaluation Results
| Method | Links | ||||||||
|---|---|---|---|---|---|---|---|---|---|
| Efficient-VLNVenue=Not Specified2025.12 | 16.2 | 28.3 | 45.1 | 30.6 | 22.5 | 46.2 | 18.7 | 95.6 | |
| Video-3D LLMVenue=Not Specified2025.12 | 16.17 | 30.1 | 47.05 | 31.7 | 22.83 | 49.02 | 19.84 | 102.06 | |
| StreamVLNframes=16, Venue=Not Specified2025.12 | 15.7 | 28.8 | — | — | — | 48.3 | 19.8 | 100.2 | |
| NaVILAframes=16, Venue=RSS20252025.12 | 15.2 | 27.4 | — | — | — | 48.3 | 19.6 | 99.8 | |
| LLaVA-3DVenue=ICCV20252025.12 | 14.5 | 27 | — | — | — | 50.1 | 20.7 | 91.7 | |
| ChatSceneVenue=NeurIPS20242025.12 | 14.31 | 21.62 | 43.2 | 29.06 | 20.57 | 41.56 | 18 | 87.7 | |
| LL3DAVenue=CVPR20242025.12 | 13.53 | — | — | — | — | 37.31 | 15.88 | 76.79 | |
| NaviLLMVenue=CVPR20242025.12 | 12.5 | 23 | — | — | — | 38.4 | 15.4 | 75.9 | |
| 3D-LLMBackbone=BLIP2-flant5, Venue=NeurIPS20232025.12 | 12 | 20.5 | 39.3 | 25.2 | 18.4 | 35.7 | 14.5 | 69.4 | |
| LEOVenue=ICML20242025.12 | 11.5 | — | — | — | — | 39.3 | 16.2 | 80 | |
| 3D-VisTAVenue=ICCV20232025.12 | 10.4 | 22.4 | — | — | — | 35.7 | 13.9 | 69.6 | |
| ScanQAVenue=CVPR20222025.12 | 10.08 | 21.05 | 30.24 | 20.4 | 15.11 | 33.33 | 13.14 | 64.86 | |
| 3D-LLMBackbone=Flamingo, Venue=NeurIPS20232025.12 | 7.2 | 20.4 | 30.3 | 17.8 | 12 | 32.3 | 12.2 | 59.2 |