Updates from Black Forest Labs
Frontier AI research lab building multimodal models for visual intelligence, including the FLUX family for image, video, audio, and action.
- Self-Supervised Flow Matching for Scalable Multi-Modal Synthesis (Project update · March 3, 2026)
The company released a research report detailing a self-supervised flow matching framework. This architecture enables the scalable synthesis of multiple modalities, including image, video, and… - FLUX 3 - Real World Models: Towards Multimodal Flow Models as the Backbone of Visual Intelligence. (Product launch · March 3, 2026)
Black Forest Labs introduced FLUX 3, a multimodal foundation model that jointly processes images, video, audio, and action. The model can generate 20-second video clips with native audio and perform… - FLUX.2: Analyzing and Enhancing the Latent Space of FLUX – Representation Comparison (Project update · November 25, 2025)
Black Forest Labs published a study on enhancing the latent space of its models. The research compares different autoencoder representations, including RAE and FLUX.2, to improve diffusion model…