#fluxschnell

[iks-iivovii]

Leonardo Phoenix (left) / Flux Schnell (right): The Challenge of Multi-Character, Multi-Limb Actions, and Hair Color Specifications

Both models seem to have improved significantly compared to SD3. However, Flux Schnell appears to understand the semantic descriptions of prompts more accurately. Leonardo Phoenix might be influenced by its preset style, making it unable to correctly depict mixed Asian faces. Additionally, its representation of Tokyo neonpunk style scenes is not as good as Flux Schnell’s.