Black Forest Labs spent the last month shipping the biggest FLUX release in its history, and this week the story split into two equally interesting tracks: the company finally made FLUX 3's video generation generally available, and one of Europe's largest media houses started putting it to work on a real children's brand. Both are worth your attention, so here is the roundup.
FLUX 3 Video Goes General Availability
After a limited early-access phase, FLUX 3's video capabilities are now generally available through the BFL API and selected partners. The model generates clips of up to 20 seconds with native audio baked in, at HD 720p or Full HD 1080p resolutions. No separate sound pipeline, no post-sync: dialogue, ambience and effects are part of the generation itself.
The standout addition is Draft Mode. A draft generation returns a fast preview of your prompt at a fraction of the cost, letting you iterate on creative directions before committing. When a draft passes muster, FLUX 3 renders the final video at full quality, keeping the same subjects, composition and motion from the approved preview. For teams producing lots of short clips, that workflow change alone is meaningful.
Black Forest Labs reports that human raters now prefer FLUX 3 for both text-to-video and image-to-video generation. In internal evaluations it beats existing state-of-the-art models in text-to-video by a solid margin and ties Seedance 2.0 in image-to-video while beating everything else. As always with vendor benchmarks, treat the numbers with some skepticism, but the direction is consistent with what independent reviewers have been seeing.
The feature set at launch includes:
- Scene and camera angle changes within a single generation
- Typography rendered naturally as part of the scene
- Multilingual dialogue with natural accents and lip-syncing
- Draft Mode for cheap, fast creative iteration
- Native audio, including 720p and 1080p output
Safety work is part of the release too. Black Forest Labs says it ran pre-release evaluations with third-party partner Cinder covering the full range of supported modalities, including non-consensual intimate imagery and child sexual abuse material. The roadmap continues with FLUX 3 Image for generation and editing, plus the open-weight FLUX 3 Dev variant that the developer community has been waiting for.
Burda's Lissy PONY Rides FLUX 3 to YouTube
The more interesting story is what happens when a big media company actually deploys this stuff. Burda Media, which is an official release partner of Black Forest Labs for FLUX 3, is bringing the Lissy PONY brand from its Blue Ocean Entertainment children's magazine to video, working with Black Forest Labs and voice AI company ElevenLabs.
Lissy PONY was already produced as an AI-assisted comic using earlier FLUX models, and this project takes the obvious next step: fully on-brand animated video. Burda's pitch is that FLUX 3 holds character fidelity across episodes and languages, so the ponies look consistent even when the production pipeline is almost entirely generative. ElevenLabs supplies the instantly recognizable character voices, which stay consistent episode after episode and market after market.
A teaser is already out. Regular episodes land this autumn in German, French and Spanish, with AI dubbing from ElevenLabs opening up additional international markets without significant extra production cost. Rebecca Gottwald, Burda Media's Chief AI Officer, says the model lets the company bring its IPs to life in fully on-brand videos without the time and cost constraints of traditional productions. Anthony Charbonnier of Black Forest Labs says the company is proud to put FLUX 3 to work on the project, and Gordian Braun of ElevenLabs notes that a brand truly comes to life through its voice, which is why the Lissy PONY characters get their own voices in three languages from the first episode.
Burda is not stopping at editorial IP. The company says it is already planning to use these models for customized, scalable brand campaigns and new product development. And in a nod to the regulatory environment, Burda commits to labeling AI-generated content in line with the EU AI Act.
The takeaway from this week's FLUX news is simple: FLUX 3 is no longer just a model to benchmark. With general availability pricing on the API and a publisher-grade deployment like Lissy PONY in the wild, the question has shifted from whether the technology can do it to who will ship the next ambitious project on top of it.
Comments