FLUX.1
By Black Forest Labs
1 is a family of text-to-image diffusion transformer models released in 2024 by Black Forest Labs — a company founded by former Stability AI researchers who created the original Stable Diffusion — known for strong prompt adherence,…
Definition
FLUX.1 is a family of text-to-image diffusion transformer models released in 2024 by Black Forest Labs — a company founded by former Stability AI researchers who created the original Stable Diffusion — known for strong prompt adherence, photorealism, and accurate text rendering within generated images.
Overview
FLUX.1 was released by Black Forest Labs shortly after its founding by researchers including Robin Rombach, one of the original creators of the Stable Diffusion architecture. The model line uses a hybrid architecture the team describes as combining diffusion transformer and flow matching techniques, a departure from the UNet-based denoising backbone used in earlier diffusion models like Stable Diffusion and Stable Diffusion XL. FLUX.1 was released in three variants with different licensing and capability trade-offs: FLUX.1 [pro] (highest quality, accessible via API), FLUX.1 [dev] (open-weight, non-commercial license, guidance-distilled from the pro model), and FLUX.1 [schnell] (open-weight under Apache 2.0, distilled for very fast few-step generation suitable for local and real-time use). This tiered approach let Black Forest Labs offer both a commercial flagship product and genuinely open, self-hostable models in the same release. FLUX.1 was widely noted for improved handling of hands, complex scene composition, and rendering legible text within images — a persistent weak point for earlier diffusion models that rely on CLIP-based text encoders — and it quickly became a popular base model for the open-source image-generation community, with broad support in tools like ComfyUI shortly after release, similar to how SDXL had been adopted previously.
Key Features
- Hybrid diffusion transformer and flow matching architecture
- Released in three variants: pro (API), dev (open, non-commercial), and schnell (open, Apache 2.0)
- Created by Black Forest Labs, founded by original Stable Diffusion researchers
- Notable improvements in text rendering and hand/anatomy accuracy within images
- Schnell variant optimized for fast, few-step local generation
- Rapid adoption across open-source image-generation tooling