Black Forest Labs Launches FLUX 3 Multimodal AI for Image and 20-Second Video Generation

FLUX 3 Multimodal AI Release

Black Forest Labs (BFL) has launched FLUX 3, expanding its FLUX model family beyond image generation. This new multimodal frontier model is trained to understand and generate images, or combined audio/video clips up to 20 seconds, from a single prompt. It also extends its underlying architecture to robotic vision and actions.

Integrated Multimodal Training

BFL states that FLUX 3 is jointly trained across image, video, and audio modalities, rather than integrating separate models. This approach is central to the company's vision of visual intelligence, aiming to connect creative generation, simulation, computer use, and robotics as applications of a single capability.

Product Lines and Availability

FLUX 3 will be offered through four product lines: FLUX 3 Video, FLUX 3 Image, FLUX 3 Action, and the upcoming open-source FLUX 3 Dev. Currently, FLUX 3 Video (with optional native audio generation) and FLUX 3 Action are available through a gated "Early Access" program. Public access via API or partners is not yet available, but FLUX 3 Image is expected to roll out in the coming weeks, followed by general availability.

Future Open-Source and Open-Weight Plans

BFL has not yet announced pricing, service-level commitments, or evaluation methodologies. While FLUX 3 is not launching with downloadable weights or an open-source license, the company plans to release faster and open-weight versions later this year. FLUX 3 Dev is described as providing open-weight access to a multimodal backbone for content creation and action prediction, a broader commitment than previous image-only FLUX Dev releases.