SD3.5 Released! (Patreon)

Video coming soon!

Stability AI: https://stability.ai/news/introducing-stable-diffusion-3-5

ComfyUI Examples: https://comfyanonymous.github.io/ComfyUI_examples/sd3/

Yes, it's finally here - SD3.5 Large! But is it any good? Short answer... I think we'll have to wait and see.

As for the best settings, it's only been a few hours so I'm experimenting, but setting the SD3 Model Sampling to something LOWER than the default 3 seems to work well. Dpm++2m is a firm favourite, and a decent number of steps looks to be 30+

Being large, you'll need a decent amount of VRAM, but offloading CLIP to CPU can help there.

Have a play and see what you think ;)

As usual, the workflow used in the video is attached to this post.