Audio deepfake detection on ADD-C 1.5s duration (test) benchmark leaderboard · SOTA2 Research

Audio deepfake detection on ADD-C 1.5s duration (test)

0.53C0 ScoreS-MGAA-MFCC

Evaluation Results

Method	Links
S-MGAA-MFCC2026.01		0.53	0.59	0.59	0.64	0.84	1.35	0.75
S-MGAA-LFCC2026.01		0.98	0.94	0.95	1.13	1.49	2.48	1.33
MGAA-MFCC2026.01		1.85	1.28	1.28	1.46	1.78	2.52	1.7
MGAA-LFCC2026.01		2	2.23	2.32	2.59	3.3	5.1	2.92
LCNN2026.01		2.05	2.32	2.32	2.32	2.63	3.22	2.47
S-MGAA-CQCC2026.01		2.13	1.96	1.96	2.21	2.75	4.5	2.59
RawNet22026.01		2.35	1.82	1.84	2.17	2.47	3.93	2.43
AASIST2026.01		2.55	1.65	1.65	1.78	2.01	2.96	2.1
RawGAT-ST2026.01		2.8	1.31	1.31	1.41	1.53	2.13	1.75
MGAA-CQCC2026.01		3.73	4.08	4.05	4.48	5.18	7.5	4.84
AASIST-L2026.01		3.95	3.33	3.33	3.52	3.93	5.07	3.86
FC-LFCC2026.01		6.03	7.01	7.06	7.62	8.4	10.34	7.74