A 2 billion parameter model trained on millions of anime images and hundreds of thousands of artistic works. Write a prompt. Get a finished illustration in 30 seconds.
Run it now on Floyo!😀
Why Anima Preview 3
General-purpose models produce generic anime. Anima is trained specifically on anime and artistic imagery. It understands Danbooru tags, artist styles, and scene composition at a level general models don't reach.
At 2 billion parameters it runs fast. At 30 steps it produces clean, detailed illustrations. It accepts natural language, Danbooru-style tags, or a mix of both in the same prompt.
Danbooru tags, natural language, or combined prompts
wide anime style vocabulary trained up to September 2025
supports community LoRAs from CivitAI and HuggingFace
three sampler options for different line and style results
best results at 1024x1024