Black Forest Labs launched FLUX 3 on Thursday, marking the first time the company's flagship model generates video instead of still images. The German AI lab, known for its FLUX line of image generators, trained the new system on images, video, and audio simultaneously within one shared system. This multimodal approach enables FLUX 3 to produce clips up to 20 seconds long with audio generated alongside the picture and synced to on-screen action, including dialogue, sound effects, and ambient noise. The release represents a strategic shift from pure image generation to video capabilities, with the model also powering FLUX-mimic, a robotics application already being tested by Audi on its production line for tasks like fitting flexible door seals.
FLUX 3 Delivers 20-Second Video Clips with Synced Audio
FLUX 3 produces video clips up to 20 seconds long with audio generated alongside the picture and synced to what's happening on screen. In early evaluations, human reviewers preferred FLUX 3's output over Runway Gen-4.5 in 77% of head-to-head comparisons and over Luma Ray 3.2 in 93% of comparisons. The model beat Gemini Omni and Seedance in 52% of evaluations. These preference tests involve evaluators watching two clips and selecting the one that looks and sounds more convincing, with Black Forest Labs counting how often FLUX 3 wins.
The model also demonstrates competence in generating still images, following its legacy in image generation. Black Forest Labs shared images showing FLUX 3's versatility in producing a broad variety of styles beyond photorealism. Co-founder and CEO Robin Rombach stated, "A model that only learns images can only generate images." The company's position is that learning to predict video also means learning the physics underneath it—weight, contact, timing—which is what a machine needs to move through the physical world.
Audi Tests FLUX-mimic Robotics System on Production Line
FLUX-mimic, built with Zurich-based mimic robotics, takes FLUX 3's video-prediction engine and adds a lightweight decoder—a small add-on component that translates the model's internal sense of how things move into actual robot motions. Audi is testing it on tasks like fitting flexible door seals, work that conventional automation has struggled to handle. Mimic co-founder Stephan-Daniel Gravert said, "Audi represents the kind of manufacturing partner we built FLUX-mimic for." Audi's Christoph Schneider stated the robots now "solve complex soft-body manipulation work" that older machines couldn't touch. Black Forest Labs reports the full system reacts in about 101 milliseconds, in the neighborhood of human visual reflexes.
Black Forest Labs Built FLUX 3 After Stable Diffusion Origins
Black Forest Labs was founded in August 2024 by veteran researchers who had helped build the original Stable Diffusion models at Stability AI. The company launched Flux models that beat MidJourney and outclassed Stability's Stable Diffusion 3. The open-source Flux Dev and Schnell models grabbed the "best open source image generator" title that AI artists had expected Stable Diffusion 3.5 to reclaim. FLUX 1.1 Pro topped the Artificial Analysis image arena in October, though that version wasn't open source.
Black Forest Labs released FLUX.2 in November 2025 but it wasn't as popular. The open-source crown held by the original Flux lasted until Alibaba's Z-Image Turbo dethroned it in late 2025, matching its quality on lower end consumer graphics cards. One CivitAI user wrote at the time, "This is what SD3 was supposed to be."
Dev Version Scheduled for Later in 2026
Video and Action are in early access now through APIs and select partners, with mimic robotics among them. Image generation is following "in the coming weeks," according to Black Forest Labs. The open-weight Dev version, the only tier Black Forest Labs plans to release for local use, isn't due until later in 2026.
FAQ
What video capabilities does FLUX 3 offer?
FLUX 3 produces video clips up to 20 seconds long with audio generated alongside the picture and synced to on-screen action, including dialogue, sound effects, and ambient noise. In early evaluations, human reviewers preferred FLUX 3's output over Runway Gen-4.5 in 77% of comparisons and over Luma Ray 3.2 in 93% of comparisons.
How is Audi using FLUX-mimic on its production line?
Audi is testing FLUX-mimic on tasks like fitting flexible door seals, work that conventional automation has struggled to handle. The system, built with mimic robotics, translates FLUX 3's video-prediction engine into actual robot motions with a reaction time of about 101 milliseconds. Audi's Christoph Schneider stated the robots now solve complex soft-body manipulation work that older machines couldn't touch.
When will the open-weight Dev version of FLUX 3 be available?
The open-weight Dev version, the only tier Black Forest Labs plans to release for local use, isn't due until later in 2026. Video and Action are currently in early access through APIs and select partners, with image generation following in the coming weeks.
