Multi-view Video Generation on nuScenes (val)
5.8FIDUniMLVG
Evaluation Results
| Method | Links | |||||||||
|---|---|---|---|---|---|---|---|---|---|---|
| UniMLVGFr=30, Res=384×7042026.06 | 5.8 | 36.1 | — | 31.04 | 57.7 | 0.85 | 95 | 55.6 | 27.4 | |
| InfiniVerseInput Type=Image, Multi-view Video=Yes, 3D Representation=Voxel+LiDAR, 2D-3D Aligned=Yes2026.06 | 6.4 | — | — | — | — | — | — | — | — | |
| UnisceneInput Type=BEV, Multi-view Video=Yes, 3D Representation=Voxel+LiDAR, 2D-3D Aligned=No2026.06 | 6.45 | — | — | — | — | — | — | — | — | |
| OMNIDRIVEFr=32, Res=880×12802026.06 | 8.01 | 45.75 | — | 31.15 | 59.5 | 0.87 | 97 | 53.4 | 33.7 | |
| DriveScapeRes=1024×5762026.06 | 8.34 | 76.39 | — | — | — | — | — | — | — | |
| VistaFr=25, Res=576×10242026.06 | 8.82 | 92.32 | — | 29.19 | 49.1 | 0.77 | 90.5 | 52.1 | 34.5 | |
| GAIA-2Fr=16, Res=448×9602026.06 | 9.46 | 52.3 | — | — | — | — | — | — | — | |
| RAYNOVAResolution=384×6722026.02 | 10.5 | 91 | 1.96 | — | — | — | — | — | — | |
| DiVEInput Type=Traj.+Box+FoV, Multi-view Video=Yes, 3D Representation=No, 2D-3D Aligned=N/A2026.06 | 10.68 | — | — | — | — | — | — | — | — | |
| MagicDrive-V2Fr=33, Res=848×16002026.06 | 10.89 | 64.81 | — | 30.99 | 51.7 | 0.83 | 92.1 | 50.6 | 29.5 | |
| X-SceneInput Type=Text+Box+HD Map, Multi-view Video=Yes, 3D Representation=Voxel, 2D-3D Aligned=No2026.06 | 12.77 | — | — | — | — | — | — | — | — | |
| VistaInput Type=Action, Multi-view Video=No, 3D Representation=No, 2D-3D Aligned=N/A2026.06 | 13.97 | — | — | — | — | — | — | — | — | |
| DriveDreamer-2Fr=16, Res=384×6402026.06 | 14.32 | 55.7 | — | 29.89 | 50.6 | 0.79 | 95.2 | 51.4 | 33.1 | |
| DriveDreamerResolution=256×4482026.02 | 14.9 | 341 | 0.37 | — | — | — | — | — | — | |
| DriveDreamerInput Type=Image+Action+FoV Map, Multi-view Video=Yes, 3D Representation=No, 2D-3D Aligned=N/A2026.06 | 14.9 | — | — | — | — | — | — | — | — | |
| PanaceaFr=16, Res=256×5122026.06 | 14.91 | 244 | — | 30.01 | 50.8 | 0.8 | 93.2 | 41.5 | 34.1 | |
| DelphiFr=40, Res=512×10242026.06 | 15.08 | 113.5 | — | — | — | — | — | — | — | |
| DelPhiInput Type=BEV Map, Multi-view Video=Yes, 3D Representation=No, 2D-3D Aligned=N/A2026.06 | 15.08 | — | — | — | — | — | — | — | — | |
| DrivingDiffusionResolution=512×5122026.02 | 15.8 | 332 | — | — | — | — | — | — | — | |
| Drive-WMInput Type=Image+Action+Box/Map, Multi-view Video=Yes, 3D Representation=No, 2D-3D Aligned=N/A2026.06 | 15.8 | — | — | — | — | — | — | — | — | |
| DrivingDiffusionFr=12, Res=256×5122026.06 | 15.83 | 119.42 | — | 27.42 | — | — | — | — | — | |
| X-DriveResolution=224×4002026.02 | 16 | — | 0.83 | — | — | — | — | — | — | |
| MagicDriveResolution=224×4002026.02 | 16.2 | — | 1.76 | — | — | — | — | — | — | |
| MagicDriveInput Type=Pose+Box+BEV, Multi-view Video=Yes, 3D Representation=No, 2D-3D Aligned=N/A2026.06 | 16.59 | — | — | — | — | — | — | — | — | |
| PanaceaInput Type=Image+Box+FoV Map, Multi-view Video=Yes, 3D Representation=No, 2D-3D Aligned=N/A2026.06 | 16.87 | — | — | — | — | — | — | — | — | |
| PanaceaResolution=256×5122026.02 | 17 | 139 | 0.67 | — | — | — | — | — | — | |
| BEVWorld2026.02 | 19 | 154 | — | — | — | — | — | — | — | |
| MagicDrive-3DInput Type=Traj.+Box+BEV, Multi-view Video=Yes, 3D Representation=3DGS, 2D-3D Aligned=No2026.06 | 20.67 | — | — | — | — | — | — | — | — | |
| MagicDrive-V2Input Type=Traj.+Box+BEV, Multi-view Video=Yes, 3D Representation=No, 2D-3D Aligned=N/A2026.06 | 20.91 | — | — | — | — | — | — | — | — | |
| GenADInput Type=Image+BEV map, Multi-view Video=No, 3D Representation=No, 2D-3D Aligned=N/A2026.06 | 23.9 | — | — | — | — | — | — | — | — | |
| DriveDreamer-2Input Type=Box+FoV Map, Multi-view Video=Yes, 3D Representation=HD Map+BEV, 2D-3D Aligned=No2026.06 | 25 | — | — | — | — | — | — | — | — | |
| Drive-WMFr=12, Res=192×3842026.06 | 25.88 | 122.7 | — | 26.91 | 49.2 | 0.71 | 86.3 | 44.1 | 37.9 | |
| WovoGenInput Type=Image+HD Map+Voxel, Multi-view Video=Yes, 3D Representation=HD Map+Voxel, 2D-3D Aligned=No2026.06 | 27.6 | — | — | — | — | — | — | — | — | |
| BEVControlInput Type=BEV Sketch, Multi-view Video=No, 3D Representation=No, 2D-3D Aligned=N/A2026.06 | 29.04 | — | — | — | — | — | — | — | — | |
| DriveGANFr=16, Res=256×2562026.06 | 31.79 | 502.3 | — | 24.32 | 37.1 | 0.6 | 94.4 | 43.2 | 38.8 | |
| BEVGenInput Type=BEV Map, Multi-view Video=No, 3D Representation=No, 2D-3D Aligned=N/A2026.06 | 40.48 | — | — | — | — | — | — | — | — | |
| DriveGanInput Type=Image+Action, Multi-view Video=No, 3D Representation=No, 2D-3D Aligned=N/A2026.06 | 73.4 | — | — | — | — | — | — | — | — | |
| InfiniCityInput Type=Voxel, Multi-view Video=No, 3D Representation=Voxel, 2D-3D Aligned=No2026.06 | 77 | — | — | — | — | — | — | — | — | |
| InfiniCubeInput Type=Box+HD Map, Multi-view Video=No, 3D Representation=Voxel+3DGS, 2D-3D Aligned=No2026.06 | 80.13 | — | — | — | — | — | — | — | — | |
| DiVE2026.06 | — | 94.6 | — | — | 51.8 | — | — | — | — |