Deepfake Detection on DF40 (cross-manipulation) benchmark leaderboard · SOTA2 Research

Deepfake Detection on DF40 (cross-manipulation)

98.6UniFace Detection RateGSD

Evaluation Results

Method	Links
GSD2026.03		98.6	99.5	97.5	98.8	96.6	95.7	97.8
Effort2026.03		96.2	98.3	92.6	95.7	93.6	92.6	94.8
GSD2026.03		96.1	97.9	93.7	96.1	92.1	91	94.5
VbSaT2026.03		96	98	91.6	96.4	93.7	93.1	94.8
FSgD2026.03		91.8	87.5	83	86.3	87.4	91	87.8
CLIP2026.03		91	91.6	89.8	96	88.5	80.3	89.5
ProDe2026.03		90.8	77.1	74.7	92.8	83.7	84.4	83.9
RECCE2026.03		89.8	68.3	84.8	94.9	84.8	76.8	83.2
LSDA2026.03		87.2	69.4	72.1	93.9	85.5	69.3	79.6
CLIP2026.03		86.5	87.8	85.6	92	83.9	76.3	85.4
LSDA2026.03		85.4	68.4	75.9	83.2	81	72.7	77.8
χ2-DFD2026.03		85.2	91.2	83.8	89.9	78.4	84.9	85.6
ProDe2026.03		84.5	71	73.6	86.5	78.8	77.8	78.7
RECCE2026.03		84.2	65.2	78.3	88.4	79.5	73	78.1
IID2026.03		83.9	76.6	84.4	92.7	78.9	64.4	80.2
UCF2026.03		83.1	73.1	86.2	93.7	80.9	64.7	80.3
IID2026.03		79.5	71	79	86.4	74.4	64	75.7
UCF2026.03		78.7	69.2	80	88.1	76.8	64.9	76.3
CDFA2026.03		76.5	67.4	75.4	84.8	72	76.1	75.4
CDFA2026.03		76.2	63.1	80.3	94.2	77.2	75.7	77.8
SBI2026.03		72.4	75	59.4	80.3	71.2	70.1	71.4
SBI2026.03		64.4	69	44.7	87.9	63.3	56.8	64.4