Black Forest Labs Releases Open-Source FLUX.2 Klein 4B
The new 4-billion-parameter model is a distilled version of the powerful FLUX.2 architecture, released under a commercially-friendly Apache 2.0 license.

Black Forest Labs has released FLUX.2 klein, a new 4-billion-parameter open-source model for text-to-image generation and editing. The model brings a more compact and accessible version of the company's powerful FLUX.2 architecture to the public.
The "klein" in its name, German for "small," signifies that this is a distilled model. Distillation is a process where a smaller model is trained to replicate the performance of a much larger one, resulting in a tool that is significantly faster and requires less computational power to run. This makes high-quality image generation feasible for developers and researchers without access to enterprise-grade hardware.
Open for Business
A key aspect of this release is its permissive Apache 2.0 license. Unlike models released with more restrictive terms, FLUX.2 klein can be freely used, modified, and integrated into commercial applications. This open approach encourages wider adoption and accelerates innovation by removing significant legal barriers for startups and independent creators.
By providing a potent combination of performance, efficiency, and an open license, Black Forest Labs is offering a valuable new building block for the creative AI ecosystem. The model weights and card are available now on the Hugging Face Hub.
Sources
- Visit
black-forest-labs/FLUX.2-klein-4B
Hugging Face
More in Text → Image

Microsoft's Mage-Flow packs image editing into 4B
A compact model handles both text-to-image generation and instruction-based edits at native resolution, under a permissive MIT license.
Boogu-Image-0.1 Brings Unified Multimodal to Open Source
A new Apache-licensed model family folds bilingual text-to-image generation and instruction editing into one system.

NVIDIA distills Qwen-Image for few-step generation
A DMD2-distilled build of Qwen-Image trades sampling steps for speed while keeping the original model's output profile.
0 comments
No comments yet. Be the first to weigh in.