Share your thoughts, 1 month free Claude Pro on usSee more

Pointwise Reasoning Evaluation on DeepfakeJudge Meta-Human

0.5RMSE

DeepfakeJudge-7B

Updated 1mo ago

Evaluation Results

Method	Links
DeepfakeJudge-7B 2026.02		0.5	0.25	0.96	0.95
DeepfakeJudge-3B 2026.02		0.56	0.31	0.96	0.95
GPT-4o-Mini 2026.02		0.81	0.66	0.88	0.86
Qwen-3-VL-235B-Thinking 2026.02		0.95	0.91	0.85	0.86
Qwen-3-VL-30B-Thinking 2026.02		1.04	1.07	0.85	0.83
Gemini-Flash-2.5 2026.02		1.11	1.24	0.85	0.83
Qwen-3-VL-8B-Thinking 2026.02		1.18	1.39	0.84	0.81
Qwen-3-VL-235B-Instruct 2026.02		1.26	1.58	0.81	0.77
Qwen-3-VL-8B-Instruct 2026.02		1.28	1.63	0.79	0.76
Qwen-3-VL-30B-Instruct 2026.02		1.32	1.75	0.89	0.84
Qwen-3-VL-2B-Instruct 2026.02		1.41	1.98	0.69	0.72
Qwen-3-VL-4B-Instruct 2026.02		1.53	2.34	0.6	0.61