CottonNoob constantly pursues the most comprehensive character-trained base models, focusing on producing a wide variety of characters. It excels when you want to generate the latest characters or mix new characters together — even without LoRA. However, its maximum effective resolution is lower than CottonIllustrious, and its style tends to stay within a fixed anime aesthetic.

CottonIllustrious, built on Illustrious 2.0, delivers exceptional stability even at extremely high resolutions (up to 2048×2048). While it supports fewer learned characters compared to CottonNoob, it’s ideal for users who prioritize native high-resolution outputs and clean rendering.

LinenIllustrious stands out for its strong responsiveness to abstract tags, resulting in highly creative and artistic illustrations. It produces unique and expressive imagery distinct from the Cotton series, though it supports the fewest characters and lowest resolution range. Best suited for generating original or concept-style characters.