Just a quick addendum to my previous Anima overview: It just appeared on the CivitAI online generator, you can find it at the bottom of the ecosystem selector (under "Other" with Chroma and HiDream) or choose "create" on the model page.
Some notes:
As expected it's just for text-to-image generation. No i2i of any kind is supported. This tracks since that was only ever possible with SD 1.5 and SDXL models.
You can only specifically use Anima Preview 3 at the moment. Preview 2 and more importantly any of its finetunes are not available. EDIT: I guess you'd have to bid for them on the auction, duh. Well, next week's auction closes tomorrow. DOUBLE EDIT: Anima CAT made it in the auction, but (predictably) it didn't work: you get a "Create" button but it only loads Preview 3.
Similarly, LoRAs are not supported for it at all. I very much hope this is only temporary since Anima massively benefits from style LoRAs.It was temporary, you can use LoRAs now.It's 8 Buzz for a reasonable 25 steps, which is a bit steeper than I'd hoped – as much as a Z-Image Turbo generation. But at least it's not as massively overpriced as Flux.1.
Random weird quirk: If you try to reuse an SDXL prompt for testing, don't pick one with BREAKs in it. It appears to be severely allergic to that. (P.S. I can't replicate this locally so I don't think it's a bug with Anima or its text encoder)