AI-generated video detection on GenVideo Moon Valley
98.2AccuracyNPR
Evaluation results
| Method | Links | ||||
|---|---|---|---|---|---|
| NPRTraining=10K Kinetics-400 + 10K Pika2026.05 | 98.2 | 98 | 98.2 | 99.66 | |
| NPRTraining Data=Kinetics-400 and Pika, Evaluation Protocol=Standard2025.10 | 98.2 | 98 | 98.2 | 99.66 | |
| DeMambaTraining=10K Kinetics-400 + 10K Pika2026.05 | 97.6 | 98.8 | 97.63 | 99.68 | |
| DeMambaTraining Data=Kinetics-400 and Pika, Evaluation Protocol=Standard2025.10 | 97.6 | 98.8 | 97.63 | 99.68 | |
| NSG-VDTraining=10K Kinetics-400 + 10K Pika2026.05 | 96.67 | 100 | 96.77 | 98.15 | |
| NSG-VDTraining Data=Kinetics-400 and Pika, Evaluation Protocol=Standard2025.10 | 96.67 | 100 | 96.77 | 98.15 | |
| TALLTraining=10K Kinetics-400 + 10K Pika2026.05 | 96.2 | 93.4 | 96.09 | 99.73 | |
| TALLTraining Data=Kinetics-400 and Pika, Evaluation Protocol=Standard2025.10 | 96.2 | 93.4 | 96.09 | 99.73 | |
| Skyra-RL-GenVideoModel Size=7B2025.12 | 96.09 | 97.92 | 96.16 | — | |
| CAM-VFDTraining=10K Kinetics-400 + 10K Pika2026.05 | 90.6 | 82.3 | 86.4 | 94.9 | |
| Skyra-RLModel Size=7B2025.12 | 81.71 | 65.34 | 78.13 | — | |
| STILTraining=10K Kinetics-400 + 10K Pika2026.05 | 71.7 | 43.4 | 60.53 | 99.34 | |
| STILTraining Data=Kinetics-400 and Pika, Evaluation Protocol=Standard2025.10 | 71.7 | 43.4 | 60.53 | 99.34 | |
| Skyra-SFTModel Size=7B2025.12 | 69.01 | 38.98 | 55.17 | — | |
| DeMamba2025.12 | 62.06 | 99.84 | 72.46 | — | |
| AIGVDet2025.12 | 50 | 0 | 0 | — | |
| NSG-VD2025.12 | 49.6 | 98.88 | 66.24 | — |