FLUX.1 Kontext enables in-context image generation for enterprise AI pipelines

Black Forest Laboratories (BFL), the startup founded by the creators of Popular stable diffusion modellaunched a new image generation model called Flow. 1 Context. This model generates and modifies not only photos, but also allows users to modify them with text and other images.

The company also announced its new BFL playground, where people can try BFL models before letting them lose on corporate applications.

BFL published two versions of the model: Flux.1 Kontext [pro] and flow. 1 KONTEXT [max]. A third version, flow. 1 KONTEXT [dev] will be available on the private beta version. Pro and Max versions are now available on platforms such as Cheat,, Freepik,, Lightricks,, Openart And Leonardoai. These models allow business creative teams and other developers to modify images with precision and at a faster rate.

Today, we are publishing flow. 1 KONTEXT – A series of generative flow correspondence models that allow you to generate and modify images.

Unlike traditional models of text in the image, Kontext understands both text and images as input, allowing true generation and publishing in context. pic.twitter.com/zlejguxdge

– Black Forest Labs (@bfl_ml) May 29, 2025

Flow. 1 KONTEXT can perform a generation in the context. This means that the model can be generated from a reference or a situation presented to it; It does not generate from zero.

The company declared in an article on X that four things make Kontext “Special”:

• Consistency of characters and preservation of elements through the scenes
• Local edition which “targets specific parts without affecting the rest”
• Style reference that generates scenes in existing styles, and
• Minimum latency

Developers can test use cases and play with models on the BFL play field before accessing the complete BFL API.

Pro and max models

Companies can use the PRO version for quick and iterative mounting. Users can enter both text and reference images and make local modifications. The company said Kontext [pro] Use “up to an order of magnitude faster than the previous advanced models” and is one of the first models that allows publishing on several laps.

On the other hand, flow. 1 Kontext [max] is the faster version with maximum performance. The company said that it adheres more to prompts, makes the typography readable and is consistent in changes without compromising speed.

Of course, many other image generation models can also generate photos from downloaded files. Image editor Ai de Midjourney Can use a reference image, then modify specific regions. Also Adobe‘s Fireflyto which many people who use the popular image and video platforms of Adobe have access.

Flow. 1 Context [dev]The third version of the KONTEXT family of models, is a model open at 12 billion parameters.

Generative

BFL said that flux.1 KonText is a flow model, which gives it more flexibility to accomplish the tasks mentioned above.

Flow models learn from a continuous data flow and define a path between noisy data and useful information. This differs from the diffusion, the Model architecture that underlies many models of generation of images and videos from Stability ai,, Media And even OPENAI’s Sorawho “crumbs” the data.

BFL said in a blog article that KonText models represent an progression of flow models.

“The flow models. 1 Kontext go beyond the text in the image,” said the company. “Unlike previous flow models which only allow generation -based generation, flow models. 1 KonText also include and can create from existing images. With flow. 1 KONTEXT, you can modify an input image via simple text instructions, allowing the editing of flexible and instantaneous images – no need for finetuning or complex publishing workflow. ”

In the text reference test to the image, BFL claimed the flow models. 1 KonText can compete with other models in terms of aesthetics, depending on the guests, realism and typography.

Arouse

Bfl released the Text Mod in the imageEl Flux 1.1 Pro in October of last year. It also included an API for third -party developers to integrate it into their applications.

Thanks to the BFL playground, some users have already started playing with the KonText models and declaring to be impressed.

Can we talk about how it is breathtaking that we can use a single image to generate head shots or modify images in less than 10 seconds with @bfl_ml Flow. 1 Context?

I used the professional model on the head on @reproduce And it’s not perfect, but my real images are on the left … pic.twitter.com/szfyzuvim

– Heather Cooper (@hbcoop_) May 29, 2025

Of course, he must still compete with other models of available images, especially those that have been around for a few years and that have continued to improve.

Daily information on business use cases with VB daily

If you want to impress your boss, VB Daily has covered you. We give you the interior scoop on what companies do with a generative AI, from regulatory changes to practical deployments, so that you can share information for a maximum return on investment.

Read our Privacy Policy

Thank you for subscribing. Find out more VB Newsletters here.

An error occurred.