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

Visual Quality Evaluation on EBench-18K

0.9466SRCC

ReasonEdit

Updated 23d ago

Evaluation Results

Method
ReasonEdit 2026.05	0.9466	0.8018	0.9524
ReasonEdit 2026.05	0.9272	0.765	0.9311
LMM4Edit* 2026.05	0.9136	0.7432	0.9189
LMM4Edit* 2026.05	0.9048	0.7837	0.9176
InternVL2.5 (8B)* 2026.05	0.8841	0.7571	0.8949
InternVL2.5 (8B)* 2026.05	0.8836	0.7223	0.8861
DeepSeekVL2 (small)* 2026.05	0.8778	0.75	0.8819
DeepSeekVL2 (small)* 2026.05	0.8674	0.7303	0.8665
AHIQ* 2026.05	0.8365	0.6457	0.8515
AHIQ* 2026.05	0.8183	0.6241	0.8324
Q-Align* 2026.05	0.818	0.6285	0.8014
MANIQA* 2026.05	0.805	0.6136	0.8171
TOPIQ* 2026.05	0.7936	0.6021	0.8054
CVRKD* 2026.05	0.7935	0.5991	0.8106
CVRKD* 2026.05	0.7864	0.5917	0.8081
LPIPS (alex) 2026.05	0.7395	0.5478	0.7594
LPIPS (vgg) 2026.05	0.7248	0.5326	0.743
Q-Align* 2026.05	0.7046	0.5188	0.7321
Qwen2-VL (7B) 2026.05	0.6786	0.4866	0.7041
MANIQA* 2026.05	0.6529	0.4716	0.7041
TOPIQ* 2026.05	0.632	0.6692	0.4565
SCSSIM 2026.05	0.5868	0.4149	0.5938
Qwen2-VL (7B) 2026.05	0.5478	0.3981	0.52
GMSD 2026.05	0.5272	0.3689	0.5328
SSIM 2026.05	0.4635	0.3217	0.4865
ImageReward 2026.05	0.4033	0.2779	0.4662
ImageReward 2026.05	0.3991	0.2764	0.4351
BRISQUE 2026.05	0.3423	0.236	0.3955
VQAScore 2026.05	0.3014	0.205	0.3162
NIQE 2026.05	0.2979	0.2069	0.2453
PickScore 2026.05	0.2483	0.1666	0.2889
CLIPScore 2026.05	0.2325	0.1586	0.2581
VQAScore 2026.05	0.2185	0.1444	0.2537
CLIPScore 2026.05	0.2181	0.1467	0.2243
LPIPS (alex) 2026.05	0.1832	0.1234	0.2782
NIQE 2026.05	0.1748	0.12	0.1926
LPIPS (vgg) 2026.05	0.1643	0.1101	0.1902
DIIVINE 2026.05	0.1429	0.0929	0.3555
PickScore 2026.05	0.1357	0.0874	0.2046
BRISQUE 2026.05	0.1302	0.0883	0.2095
SCSSIM 2026.05	0.064	0.0433	0.2646
DIIVINE 2026.05	0.0121	0.0083	0.2051
GMSD 2026.05	0.0099	0.0063	0.0959
SSIM 2026.05	0.0035	0.0007	0.2133