Ready To Use

Animagine XL

Create anime images with a simple text prompt

Animagine XL 3.0 is a state-of-the-art, open-source anime text-to-image diffusion model. It represents the latest iteration in the Animagine series, building upon the capabilities of its predecessor, Animagine XL 2.0. Developed by Cagliostro Research Lab, this model is grounded on the SDXL Base framework and showcases significant advancements in anime-style image generation.

Technical Details

  • Training: Conducted on 2x A100 GPUs with 80GB memory for over 500 hours.
  • Base Model: Finetuned from Animagine XL 2.0.
  • Resolution Support: Supports various aspect ratios, including 1:1, 9:7, 7:9, 19:13, 13:19, 7:4, 4:7, 12:5, and 5:12.

Key Enhancements

  • Image Quality: Notable improvements in hand anatomy, efficient tag ordering, and enhanced knowledge about anime concepts.
  • Prompt Interpretation: Advanced prompt interpretation for superior image generation.
  • Special Tags: Includes quality, rating, and year modifiers to steer results towards desired aesthetics and themes.

Usage Guidelines

  1. Prompt Structure: For optimal results, use structured prompts with tags like character name, series, and additional details.
  2. Quality Modifiers: Use tags like 'masterpiece', 'best quality', and negative prompts (e.g., 'nsfw', 'lowres') for high-aesthetic images.
  3. CFG Scale and Samplers: Recommended CFG Scale of 5-7, sampling steps below 30, and Euler Ancestral (Euler a) sampler for best results.

Limitations and Ethical Considerations

  • Concept Over Artstyle Focus: The model emphasizes learning concepts over specific art styles, which might affect aesthetic appeal.
  • Content Sensitivity: High-scored datasets often contain NSFW content, requiring careful use of quality modifiers.


Animagine XL 3.0 is an invaluable tool for artists, illustrators, and creatives, enabling the transformation of text descriptions into high-quality anime-style images. Its sophisticated technology and enhancements make it a significant asset in character concept development and scene design.