The open source image generator Sable Diffusion has been updated to version 3.5 and fixes some bugs.

Stable Diffusion has been updated to version 3.5. The new AI image generator model processes input prompts more accurately than other generators and rivals many major models in output quality. Additionally, the output now reflects a greater variety of styles and skin tones without the user having to explicitly ask.

Three models

The most powerful of the three devices is Stable Diffusion 3.5 Large, which offers the highest quality and best-in-class user prompt tracking. The Large model is intended for professional use with a resolution of 1 MP.

Then there is the lightweight version of the Large model called Large Turbo, which focuses more on efficiency than maximum quality. Finally, there is the medium model, which is designed to run on consumer hardware. Quality and simplicity are in perfect balance. You can easily produce images with a resolution between 0.25 and 2 megapixels. The last model won’t be released until October 29th, the first two are already available.

In April, Stable Diffusion introduced their first video model, which can turn a video from eight different perspectives into new videos. This model has been available on Hugging Face since June.