Audio-Visual Deepfake Detection
Benchmarks
DatasetSOTA methodMetricTrendResultsLast updated
99.91Accuracy
17
May 1, 2026
97.68ACC (Audio-Visual)
16
Jun 1, 2026
96.5Accuracy
11
Feb 26, 2026
100AUC
10
Feb 26, 2026
85.7AP
7
May 28, 2026
95.6AP
7
May 28, 2026
85.3AP
7
May 28, 2026
92.82Accuracy
6
May 1, 2026
0.835AUC
5
Feb 26, 2026
99.2AUC
5
Feb 26, 2026
7.1EER (zero-shot)
4
May 4, 2026
5.8EER (zero-shot)
4
May 4, 2026
7.1EER
4
May 4, 2026
5.8EER
4
May 4, 2026