Black Forest Labs makes FLUX 3 Video generally available

Black Forest Labs has made FLUX 3 Video generally available through its own API and select partners. The new model generates video clips up to 20 seconds long natively in HD, with Full HD available via upscaling — and it ships with a soundtrack built in.

The standout feature is native audio generated alongside the video itself: multilingual dialogue with natural accents, sound effects, ambient noise, and lip-sync across more than 13 languages, including English, Chinese, Spanish, French, German, Japanese, Portuguese, Russian, Italian, Indonesian, Turkish, Hindi, and Punjabi. Per Black Forest Labs' own internal evaluation, the model scored a 1135 ELO rating in text-to-video generation and, the company says, "outperforms existing SOTA models... by a solid margin" — while in image-to-video, FLUX 3 ties Seedance 2.0 and Sora.

• Clips up to 20 seconds long, native HD, up to Full HD via upscaling
• Built-in audio: dialogue, effects, ambience, and lip-sync in 13+ languages
• 1135 ELO rating in text-to-video, per the company's internal benchmarks

The comparison to Seedance 2.0 lands especially hard given that ByteDance's international rollout of Seedance 2.0 has been paused since March 2026 over copyright claims from Disney, Paramount, and Netflix, who accused the company of training the model on their characters without permission. While one of its biggest rivals is stuck in legal limbo, Black Forest Labs is entering the generative-video market with a shipping, API-accessible product — part of a broader wave of corporate bets on applied AI we've already seen, for instance, in how MacPaw is using Liquid AI to build a local AI stack on Mac. But not every generative AI product ships without controversy: just days ago, Google had to pull its own AI satellite-image generator from Google Earth less than a day after launch, after users started generating fake imagery of real locations with it.

None of this should be read as personalized investment advice.