Stable Diffusion 3.5 follows your prompts more closely and generates more diverse people
Briefly

The 3.5 model tries to right some of the wrongs of the widely panned Stable Diffusion 3 Medium, improving prompt adherence and output quality.
Stability AI's new model, particularly the 3.5 Large, is designed for professional use at 1 MP resolution, highlighting its enhanced functionality.
The 3.5 series has new filters aimed at better reflecting diverse human features, addressing previous shortcomings in representation.
Stable Diffusion 3.5 Medium is aimed at consumer hardware, balancing quality and ease of use with customizable image generation capabilities.
Read at Engadget
[
|
]