Anima Turbo & Aesthetic v1.0: Fast Anime Image Models
ComfyUI Wiki
Anima Turbo and Aesthetic v1.0 are CircleStone Labs' distilled and quality-tuned variants of the 2B anime text-to-image model, natively supported in ComfyUI.
A
Anima Turbo and Aesthetic v1.0
Text-to-ImageAnimeArtisticTurboDistilledAnima Turbo and Aesthetic v1.0 are the official few-step and quality-tuned variants of Anima, CircleStone Labs' 2 billion parameter anime text-to-image model. Turbo is distilled for fast 8 to 12 step generation at CFG 1, while Aesthetic is fine-tuned on high-quality images only for a stronger default art style. Both share the Qwen 3 0.6B text encoder and Qwen Image VAE of the Anima family and run natively in ComfyUI.
| Developer | CircleStone Labs (with Comfy Org) |
| Release Date | 2026-07 |
| Base Model | Anima (Cosmos-Predict2-2B-Text2Image finetune) |
| Architecture | ~2B parameter diffusion transformer (DiT) |
| License | CircleStone Labs Non-Commercial |
| Text Encoder | Qwen 3 0.6B |
| VAE | Qwen Image VAE |
| Resolution | 512 to 1536 px |
| Turbo Settings | 8-12 steps at CFG 1 |
| Aesthetic Settings | 30-50 steps at CFG 4-5 |
Guides and workflows related to this model series.
No articles found.
Comments
Sign in with GitHub to join the discussion.