Black Forest Labs has unveiled FLUX 1 Kontext, dated 29 May 2025. The release pairs two in-context image models—Kontext [pro] and the experimental Kontext [max]—with a web Playground that runs the same API. Kontext adds joint text-and-image prompting to the FLUX line, letting users keep a reference photo and request edits or brand-new scenes in one loop.
Today we're releasing FLUX.1 Kontext - a suite of generative flow matching models that allow you to generate and edit images.
— Black Forest Labs (@bfl_ml) May 29, 2025
Unlike traditional text-to-image models, Kontext understands both text AND images as input, enabling true in-context generation and editing. pic.twitter.com/zleJGuXDge
Each model builds on a 12 B-parameter rectified-flow transformer distilled to sample in as few as eight steps. Internal tests show up to 8× lower latency than FLUX 1.1 Pro and top scores on the new KontextBench for text editing and character preservation. Features include local edits, style transfer, on-image text rewrite, and stable character identity; the “max” variant also raises typography quality and prompt following.
Kontext [pro] and [max] are live today through the BFL API plus partner services such as KreaAI, Freepik, Lightricks, Leonardo, Replicate, FAL, Runware, and Together. An open-weight 12 B Kontext [dev] enters private beta, with wider release planned on Hugging Face after safety review.
Builders using the Replicate endpoint report crisper edits at lower cost than OpenAI’s gpt-image-1 and praise consistent subjects during multi-step workflows. For Black Forest Labs—formed by alumni of the original Stable Diffusion project—Kontext deepens its hybrid strategy: commercial high-speed endpoints paired with research weights the community can fine-tune.
