AI-generated image detection on UniversalFakeDetect

100Pro-GAN AccuracyUniFD

Evaluation results

Method
UniFD2025.08	100	98.5	94.5	82	99.5	97	66.6	63	57.5	59.5	72	70.03	94.19	73.76	94.36	79.07	79.85	78.14	86.78	81.38
RINE2025.08	100	99.3	99.6	88.9	99.8	99.5	80.6	90.6	68.3	89.2	90.6	76.1	98.3	88.2	98.6	88.9	92.6	90.7	95	91.31
MiraGe2025.08	100	94.3	96.5	96.8	93.6	96.1	88.7	75.8	71.9	92.9	92.9	82	98.3	94.6	98.6	97.5	97.5	98	98.6	92.9
IAPL2025.08	100	98.6	98.65	94.89	99.39	96.7	95.89	90.83	93.84	92.47	92.72	72.75	99.5	97.7	99.15	97.95	98.3	98.35	98.9	95.61
UniFD2026.03	100	98.5	94.5	82	99.5	97	66.6	63	57.5	59.5	72	70.03	94.19	73.76	94.36	79.07	79.85	78.14	86.78	81.38
UniFD2026.04	100	98.5	94.5	82	99.5	97	66.6	63	57.5	59.5	72	70.03	94.19	73.76	94.36	—	79.85	78.14	86.78	81.38
AIDE2025.08	99.99	98.49	83.95	99.65	73.24	99.9	54.09	68.06	71.14	60.87	61	88.5	98.2	97.45	98.4	98.15	98.4	97.85	97.5	86.57
CNN-Spot2026.03	99.99	85.2	70.2	85.7	78.95	91.7	53.47	66.67	48.69	86.31	86.26	60.07	54.03	54.96	54.14	60.78	63.8	65.66	55.58	69.58
CoRIT2026.03	99.99	99.7	98.95	93.33	99.2	100	51.9	87.22	59.36	97.35	97.35	78.25	97	86.6	96.55	88.9	91.25	88.65	95.35	89.84
CNN-Spot2026.04	99.99	85.2	70.2	85.7	78.95	91.7	53.47	66.67	48.69	86.31	86.26	60.07	54.03	54.96	54.14	—	63.8	65.66	55.58	69.58
C2P-CLIP2025.08	99.98	97.31	99.12	96.44	99.17	99.6	93.77	95.56	64.38	93.29	93.29	69.1	99.25	97.25	99.3	95.25	95.25	96.1	98.55	93.79
REVEAL2026.04	99.98	97.35	99.15	96.29	99.26	99.8	93.49	95	64.16	95.34	95.34	68.9	99.25	96.95	99.3	—	95.2	95.95	98.45	93.9
SPARK-IL2026.04	99.92	97.6	99.2	96.8	99.4	99.1	94.3	95.8	65.2	96.1	96	69.8	99.4	96.1	99.6	—	95.9	96.2	98.9	94.6
FatFormer2025.08	99.89	99.32	99.5	97.15	99.41	99.75	93.23	81.11	68.04	69.45	69.45	76	98.6	94.9	98.65	94.35	94.65	94.2	98.75	90.86
SAFE2025.08	99.86	98.87	89.73	98.04	91.52	99.9	93.1	85.56	95.91	50.1	50.1	82.35	98.8	98.65	98.8	95.75	96.6	97.25	97.5	90.44
LGrad2025.08	99.84	85.39	82.88	94.83	72.45	99.62	58	62.5	50	50.74	50.78	77.5	94.2	95.85	94.8	87.4	90.7	89.55	88.35	80.28
NPR2025.08	99.84	95	87.55	96.23	86.57	99.75	76.89	66.94	98.63	50	50	84.55	97.65	98	98.2	96.25	97.15	97.35	87.15	87.56
LGrad2026.03	99.84	85.39	82.88	94.83	72.45	99.62	58	62.5	50	50.74	50.78	77.5	94.2	95.85	94.8	87.4	90.7	89.55	88.35	80.28
NPR2026.03	99.84	95	87.55	96.23	86.57	99.75	76.89	66.94	98.63	50	50	84.55	97.65	98	98.2	96.25	97.15	97.35	87.15	87.56
LGrad2026.04	99.84	85.39	82.88	94.83	72.45	99.62	58	62.5	50	50.74	50.78	77.5	94.2	95.85	94.8	—	90.7	89.55	88.35	80.28
F3Net2026.03	99.38	76.38	65.33	92.56	58.1	100	63.48	54.17	47.26	51.47	51.47	69.2	68.15	75.35	68.8	81.65	83.25	83.05	66.3	71.33
F3Net2026.04	99.38	76.38	65.33	92.56	58.1	100	63.48	54.17	47.26	51.47	51.47	69.2	68.15	75.35	68.8	—	83.25	83.05	66.3	71.33
AntiFakePrompt2026.04	99.26	96.82	87.88	80	98.13	83.57	60.2	70.56	53.7	79.21	79.01	73.75	89.55	64.1	89.8	—	93.9	92.95	80.1	82.42
Bi-LoRA2026.04	98.71	96.74	81.18	78.3	96.3	86.32	57.78	68.89	52.28	73	82.6	65.1	85.15	59.2	85	—	85.65	84.9	72.7	78.59
FreqNet2025.08	97.9	95.84	90.45	97.55	90.24	93.41	97.4	88.92	59.04	71.92	67.35	86.7	84.55	99.58	65.56	85.69	97.4	88.15	59.06	85.09
FreqNet2026.03	97.9	95.84	90.45	97.55	90.24	93.41	97.4	88.92	59.04	71.92	67.35	86.7	84.55	99.58	65.56	85.69	97.4	88.15	59.06	85.09
Co-occurrence2026.03	97.7	63.15	53.75	92.5	51.1	54.7	57.1	63.06	55.85	65.65	65.8	60.5	70.7	70.55	71	70.25	69.6	69.9	67.55	66.86
Co-occurrence2026.04	97.7	97.7	53.75	92.5	51.1	54.7	57.1	63.06	55.85	65.65	65.8	60.5	70.7	70.55	71	—	69.6	69.9	67.55	66.86
PatchForensics2026.03	75.03	68.97	68.47	79.16	64.23	63.94	75.54	75.14	75.28	72.33	55.3	67.41	76.5	76.1	75.77	74.81	73.28	68.52	67.91	71.24
Patchfor2026.04	75.03	68.97	68.47	79.16	64.23	63.94	75.54	75.14	75.28	72.33	55.3	67.41	76.5	76.1	75.77	—	73.28	68.52	67.91	71.24
Freq-spec2026.03	49.9	99.9	50.5	49.9	50.3	99.7	50.1	50	48	50.6	50.1	50.9	50.4	50.4	50.3	51.7	51.4	50.4	50	55.45
Freq-spec2026.04	49.9	99.9	50.5	49.9	50.3	99.7	50.1	50	48	50.6	50.1	50.9	50.4	50.4	50.3	—	51.4	50.4	50	55.45