What models do you prefer, and how are you running them?
I personally prefer Stable Diffusion 1.4, sometimes 1.5 because I enjoy generating Sci-fi settings and mostly landscapes rather than portraits. Also I run it on my own machine (32GB of ram and an RTX 3070ti)
Ive been using Revanimated v122
seems to work really well.
I mostly use the DreamShaper model. Very easy to get good images from more specific prompts. https://civitai.com/models/4384/dreamshaper
I have tried many from civitai, it is amazing what changing models can do. Also, find a good set of popular artists and amend prompts with: "by George Lucas", or "by Don bluth". It is like downloading a completely different model.