20s on 20 (50?) - Moxie Diffusion XL v1.6 Lightning

I test new AI models by using a lot of the same prompts. This model isn't new, but I forgot to do it when it came out. This particular prompt has its own tag. Using the same prompt isn't necessarily the best comparison when some models favor different wording styles, but it can give you an idea of what works well and what doesn't.

The last three samples I published were grouped that way because they're all updates that try to make the largely archaic SDXL architecture look better than it should. Moxie, by name at least, seems tailor-made to do well in this test. And, although trapped in the limitations that come with SDXL (women almost always looking like a clone of each other is a big one), it does pretty well and manages to have a slightly different take on it. But I really don't get why so many of these read "art-deco barrette" and put the same flower in her hair.

For other images in this project, see the tag: /tag/the-20s-on-20