Loading...
Loading...
Founded in 2024 by Robin Rombach, Andreas Blattmann and Patrick Esser, who previously worked on Latent Diffusion and Stable Diffusion, Black Forest Labs is headquartered in Freiburg, Germany, with a second HQ in San Francisco. Its FLUX.1 line (August 2024) is a 12B-parameter text-to-image family with API [pro], open-weight [dev] and Apache 2.0 [schnell] tiers, extended by FLUX 1.1 [pro] Ultra and the FLUX.1 Kontext editing models. FLUX.2 (November 2025) added multi-reference editing across [pro], [flex], [dev] and [klein] tiers, and FLUX 3 (July 2026) is a multimodal foundation model whose first product, FLUX 3 Video, reached general availability in August 2026. The lab closed a $300M Series B in December 2025.
12B rectified-flow transformer, open weights for non-commercial use
four times the resolution (up to 4MP) plus Raw mode
higher prompt adherence and typography for editing
in-context image editing from text and reference images
Advanced context-aware generation with style consistency.
SRPO-tuned FLUX.1 [dev] variant
managed API tier, editing from up to 10 reference images
32B open-weight model, generation and multi-reference editing in one model