FLUX

Black Forest Labs develops FLUX models for visual intelligence across image, video, audio, and action prediction. Its models are designed to understand, reason about, and act in the world, with FLUX 3 presented as a multimodal model.

The models can be accessed through a browser playground, an API, or open-weight deployments on a user's own infrastructure. The site describes support for text, images, and keyframes as inputs, with image generation, video clips of up to 20 seconds, generated audio, and robotics-oriented visual and control predictions.

The API is positioned for production workloads, while open-weight access supports deployment, fine-tuning, and customization. Enterprise options, API pricing, documentation, and licensing are available through the site.