Black Forest Labs releases FLUX 3 Video, tops text-to-video rankings

Black Forest Labs announced the general availability of FLUX 3 Video, its latest model for video generation, accessible through the BFL API and select partners. The model can generate HD and Full HD video clips up to 20 seconds in length, each with native audio containing dialogue, sound effects, and ambient noise.

FLUX 3 supports multiple workflows: text-to-video conversion, image-to-video generation, keyframe-based control, video continuation, and the ability to include multiple scenes and camera angles within a single output clip. A distinctive feature is its capacity to render text and typography directly into scenes, follow complex user prompts, and apply world knowledge—enabling applications such as documentary production. The model also generates lip-synced dialogue in more than 14 languages, broadening its appeal for international content creation.

According to Black Forest Labs' internal Elo rankings—a comparative scoring system for model outputs—FLUX 3 Video achieves a score of 1,135 in text-to-video generation and 1,051 in image-to-video generation, rankings that place it ahead of Gemini Omni Flash, Minimax H3, and Seedance 2.0. Pricing operates on a per-second-of-output basis. In draft mode (HD quality only), text-to-video and image-to-video generation costs $0.06 per second, while video-to-video costs $0.12 per second. Full quality HD pricing is $0.17 per second for text-to-video or image-to-video, and $0.41 for video-to-video. Full HD quality costs $0.29 per second for text-to-video or image-to-video, and $0.53 for video-to-video. Audio is included at all pricing tiers. Additional video examples are available on the Black Forest Labs blog.