Black Forest Lab FLUX3 Multimodal Model Makes Its Debut: Generate 20-Second Audio and Video in One Go, Significantly Outperforming Grok and Seedance

Black Forest Labs announced the launch of FLUX3, a multi-modal foundation model, in Early Access mode. The model adopts a unified architecture for joint learning of images, videos, and audio. Based on the Self-Flow (self-supervised flow matching) learning framework, it synchronously trains videos, images, and audio, further expanding from the FLUX.1 and FLUX.2 series to multi-modal generation and understanding tasks.

Completely superior performance, native audio synchronization output

FLUX3 can output videos up to 20 seconds long in a single generation, along with native audio. It supports various creation modes such as text-to-video, image-to-video, video-to-video, input video and audio continuation, keyframe-to-video, multilingual dialogue, and multi-shot video concatenation. In manual evaluation of 10-second 720p videos with sound, FLUX3 achieved a win rate of 69% against Grok Imagine Video, and 52% against Seedance 2.0 and Gemini Omni Flash, demonstrating a comprehensive leading advantage.

Comprehensive image capabilities, entering robot behavior prediction

Black Forest Labs releases Flux3: The first native multimodal foundation model for generating audio 20 seconds of synchronized audio-visual content at a time

Black Forest Lab releases Flux3, a multimodal model built on Self-Flow architecture, integrating image, video, audio, and motion codecs for unified understanding and generation of physical and digital worlds. It is the first to support native audio generation, producing 20-second synced audio-video clips in one pass. Capabilities include text, image, and video-to-video tasks.....

Germany Black Forest Laboratory Releases Flux3 Multimodal Model: Native Audio Generation 20-Second Synchronized Video Output

Black Forest Labs unveils Flux3, a multimodal foundation model using Self-Flow architecture with dedicated encoders/decoders for image, video, audio, and motion. It unifies understanding and generation across physical and digital environments, outperforming Luma and Runway. Flux3 natively generates audio, outputs 20s sync’d video with audio, and supports text/image/video-to-video, keyframe transitions, and multilingual dialogue.....

Black Forest Lab FLUX3 Multimodal Model Makes Its Debut: Generate 20-Second Audio and Video in One Go, Significantly Outperforming Grok and Seedance

Related Recommendations

Black Forest Labs releases Flux3: The first native multimodal foundation model for generating audio 20 seconds of synchronized audio-visual content at a time

Germany Black Forest Laboratory Releases Flux3 Multimodal Model: Native Audio Generation 20-Second Synchronized Video Output

Black Forest Labs Releases FLUX.2 Model with a 66.6% Win Rate Leading Open-Source Benchmarks

Developer Version of Stable Diffusion! FLUX. 2 Open Source Release: 10 Image References + 4MP Editing

Bid Farewell to AI-like Aesthetics! Black Forest Labs Collaborates with Krea to Open-source FLUX.1-Krea Model Redefining Natural Aesthetics in Image Generation