AI-generated video detection on GenVideo ModelScope
98.33AccuracyNSG-VD
Evaluation results
| Method | Links | ||||
|---|---|---|---|---|---|
| NSG-VDTraining=10K Kinetics-400 + 10K Pika2026.05 | 98.33 | 0.9833 | 0.9833 | 0.9866 | |
| DeMambaTraining=10K Kinetics-400 + 10K Pika2026.05 | 95 | 0.936 | 0.9493 | 0.9882 | |
| CAM-VFDTraining=10K Kinetics-400 + 10K Pika2026.05 | 94.8 | 0.895 | 0.921 | 0.972 | |
| NPRTraining=10K Kinetics-400 + 10K Pika2026.05 | 89.2 | 0.8 | 0.8811 | 0.9718 | |
| ReConFuseModality=Video, Training Required=Yes2026.06 | 87.7 | 0.892 | 0.8788 | 0.958 | |
| DeMambaModality=Video, Training Required=Yes2026.06 | 86.7 | 0.742 | 0.848 | 0.9631 | |
| STILTraining=10K Kinetics-400 + 10K Pika2026.05 | 85.4 | 0.708 | 0.829 | 0.9777 | |
| TALLTraining=10K Kinetics-400 + 10K Pika2026.05 | 82.1 | 0.652 | 0.7846 | 0.9714 | |
| Skyra-RL-GenVideoModel Size=7B2025.12 | 79.93 | 0.6629 | 0.7676 | — | |
| DIREModality=Image, Training Required=Yes2026.06 | 79.8 | 0.612 | 0.7518 | 0.9305 | |
| AEROBLADEModality=Image, Training Required=No2026.06 | 78.8 | 0.586 | 0.7343 | — | |
| D3Modality=Video, Training Required=No2026.06 | 64.5 | 0.925 | 0.718 | — | |
| DeMamba2025.12 | 62 | 0.9643 | 0.7132 | — | |
| Skyra-RLModel Size=7B2025.12 | 57.21 | 0.1671 | 0.2809 | — | |
| Skyra-SFTModel Size=7B2025.12 | 52.71 | 0.0614 | 0.115 | — | |
| AIGVDet2025.12 | 50.36 | 0.0071 | 0.0142 | — | |
| NSG-VD2025.12 | 49.79 | 0.9914 | 0.6638 | — |