German AI startup Black Forest Labs, known for powering image generation in X’s Grok chatbot, has released a new family of text-to-image models under the name Flux.1 Kontext. The announcement marks the company’s most significant release to date in the escalating race for next-gen visual AI dominance.
Leading the lineup is Flux.1 Kontext [pro], which can generate images from text and reference photos with exceptional prompt accuracy, photorealism, and typography. The model supports multi-turn refinement while preserving visual consistency. Flux.1 Kontext [max] prioritizes speed and precision, delivering results up to eight times faster than current state-of-the-art systems.
“Our new models achieve industry-leading quality while dramatically improving inference performance,” the company said in a blog post.
The models arrive as competition intensifies, with recent launches from Google’s Imagen 4 and OpenAI’s viral image tools inside ChatGPT. While Flux.1 Kontext [pro] and [max] are cloud-based only, a research-focused version, Flux.1 Kontext [dev], is available via private beta for safety testing.
Users can access the models directly through Black Forest Labs’ new model playground, which includes 200 free credits at sign-up. The company, whose founders include former Stability AI researchers, has attracted backing from Andreessen Horowitz, Oculus co-founder Brendan Iribe, and Garry Tan of Y Combinator. Reports last year indicated the startup was in talks to raise $100 million at a $1 billion valuation.
Featured image: Credit: Black Forest Labs