Black Forest Labs Launches FLUX 3 Multimodal Video Generation Model
Black Forest Labs has officially launched FLUX 3, a multimodal AI model capable of generating up to 20 seconds of native 1080p video with synchronized audio. The model utilizes a unified architecture to jointly learn from images, videos, and audio. FLUX 3 outperforms major competitors like ByteDance's Seedance 2.0 and MiniMax's H3 in text-to-video and image-to-video benchmarks. This release intensifies competition in the generative video space by offering high-quality, native audio-video generation. The model supports text-to-video, image-to-video, video-to-video, and multi-shot sequencing, with pricing ranging from $0.06 per second for Draft HD to $0.53 per second for Full HD video-to-video generation.
## BACKGROUND
Black Forest Labs is a prominent AI research organization known for developing the FLUX series of image generation models. Traditionally, AI video generation required separate models for visual rendering and audio synthesis, but modern multimodal architectures like FLUX 3 and MiniMax H3 generate both elements simultaneously in a single pass.