lux Chroma is an 8.9 billion-parameter, open-source, text-to-image foundational model based on FLUX.1-schnell (the speedy variant of the Flux model family). Licensed under Apache 2.0, it’s fully free to use, adapt, and build upon—no corporate restrictions, no gatekeeping.
Fully uncensored: Flux Chroma reintegrates content previously subject to censorship or removal—such as certain anatomical elements—making it a powerful tool for creative freedom
Ideal base for fine-tuning: As a foundational model, Chroma offers a neutral but powerful starting point for researchers, artists, and developers to experiment and expand upon
Why Flux Chroma Matters
Built on top of FLUX.1-schnell, Flux Chroma benefits from a 12B parameter rectified-flow transformer architecture optimized for rapid inference and high-quality outputs. Flux.1-schnell is renowned for delivering professional-grade images in just 1–4 inference steps—a boon for real-time or rapid-iteration workflows.
Flux Chroma’s training set includes around 5 million examples curated from a larger pool of 20 million spanning diverse content types—anime, furry art, photography, and more—making it truly versatile and uncensored
Performance & Use Cases
A model famed for matching closed-source alternatives like Midjourney v6 in quality while remaining open-source.
Chroma excels at generating rich, varied images quickly, even on modest hardware. Whether you’re prototyping, creating stylized art, or developing custom AI tools, Chroma offers great flexibility and performance.
You can now run two versions of Chroma directly inside the Sogni apps: Chroma v.48 [detail calibrated]: tuned for refined detail and higher fidelity. Chroma v.46 [flash]: optimized for speed and responsive iteration.
We use cookies to enhance your browsing experience and analyze site traffic. By clicking 'Accept,' you agree to our use of cookies. Manage your preferences anytime.
