Black forest labs unveils ultra-fast open-source AI image model FLUX.2 [klein]

German AI startup Black Forest Labs (BFL) has launched FLUX.2 [klein], a new open-source image generation model designed to create AI images in under a second, marking a push toward faster and more accessible generative tools for developers and enterprises.

Founded by former Stability AI engineers, BFL said the latest release focuses on speed, efficiency and lower computing demands, enabling near-instant image generation even on consumer-grade hardware. The FLUX.2 [klein] family includes two models — a 4-billion-parameter version released under the Apache 2.0 license, allowing full commercial use, and a larger 9-billion-parameter model restricted to non-commercial use.

Unlike its earlier high-end models aimed at photorealism, FLUX.2 [klein] is built for latency-critical workflows such as real-time design, gaming and rapid prototyping. According to the company, the 4B model can generate or edit images in less than half a second and comfortably runs on widely used GPUs, including Nvidia’s RTX series.

The models support text-to-image generation, advanced image editing and multi-reference composition within a single architecture, reducing the need for multiple tools. Features such as precise color control using hex codes and structured prompts are aimed at professional design and enterprise automation.

Industry analysts say the open licensing of the 4B model positions FLUX.2 [klein] as a strong competitor to other open-weight image generators, particularly for startups seeking low-cost, locally hosted AI solutions. With official integrations for tools like ComfyUI already available, early user feedback has highlighted the model’s speed as its key advantage.

The release reflects a broader shift in generative AI toward practical deployment, where speed, cost efficiency and data control are increasingly critical — especially for companies looking to run AI systems on their own infrastructure rather than relying on external platforms.