Semantic-aware image editing on Semantic-aware image editing dataset (test) benchmark leaderboard · SOTA2 Research

Semantic-aware image editing on Semantic-aware image editing dataset (test)

31.62PSNRFLUX.1-Fill-dev

Evaluation Results

Method	Links
FLUX.1-Fill-dev2026.04		31.62	0.9452	0.9883	0.1919	0.2055	0.074	0.3798	-1.3512	3.9582	0.0264	0.3366
Ours-ρ-UserMask2026.04		25.67	0.8758	0.9551	0.2112	0.2458	0.1013	0.6953	-0.2284	3.9603	0.0457	0.3823
Ours-ρ-AutoMask2026.04		25.31	0.8534	0.9304	0.1958	0.2457	0.076	0.657	-0.2343	3.9602	0.0304	0.379
ICEdit2026.04		25.12	0.844	0.8418	0.2046	0.2194	0.0773	0.3635	-0.913	3.9601	0.0392	0.347
DiffEdit2026.04		24.13	0.8397	0.9276	0.1664	0.1805	0.0002	0.0443	-1.9112	3.9596	0.0009	0.2952
Nano Banana2026.04		23.81	0.8163	0.9683	0.1973	0.1987	0.1088	0.6595	0.2074	3.9592	0.0319	0.3166
FLUX.1-Kontext-dev2026.04		15.39	0.4575	0.6764	0.2078	0.2214	0.0748	0.7486	0.799	3.9577	0.0424	0.2877