FLUX 3 Image to Video

Animate images into 20-second clips with synchronized native audio.

Example output

Flux 3 Image to Video

Flux 3 Image to Video is Black Forest Labs' image-to-video model that turns one or more still frames into cinematic clips with synchronized native audio.

What is Flux 3 Image to Video?

Flux 3 Image to Video is the image-to-video mode of FLUX 3, Black Forest Labs' first unified multimodal model, jointly trained across image, video, and audio. You provide keyframes and a prompt, and the model animates them into a coherent clip of up to 20 seconds with dialogue, sound effects, and ambience generated in the same pass. Because motion, physics, and sound are learned together, impacts land on the right frame and objects carry believable mass and momentum. It is BFL's first video-capable release and the successor to the image-only FLUX.1 and FLUX.2 families.

Key Features

  • Image-to-video from 1 to 10 keyframes: one image animates a still, two set first and last frame, and more interpolate in order like a storyboard.
  • Native synchronized audio, on by default, including multilingual dialogue with lip-sync, sound effects, and ambient sound.
  • HD and Full HD output at 24 fps, up to 20 seconds, across eight aspect ratios from 21:9 to 9:16.
  • Draft mode for fast, low-cost previews before committing to a full-quality render.
  • Physically grounded motion, strong facial expressions, and accurate in-frame text and typography.

Best Use Cases

Flux 3 Image to Video is ideal for animating product shots, concept art, and photographs into short cinematic sequences. In testing, a single landscape still produced a clean HD clip with a slow camera push-in, rolling waves, drifting clouds, and matching ambient audio. Use it for social verticals, ad creative, storyboard previews, music-driven visuals, and multi-shot narrative clips where a subject stays consistent.

Prompt Tips and Output Quality

Describe the subject, the camera move, the motion, and one line about sound. Pin first and last frames for precise arcs, or stack keyframes for longer shots. Testing showed stable geometry and no visible artifacts across repeated runs at HD with audio enabled.

FAQs

How many images can I use as keyframes? One to ten. One animates, two pin first and last frame, and more interpolate in sequence.

Does it generate sound? Yes. Native audio is on by default and can be turned off for silent clips.

What resolutions and lengths are supported? HD or Full HD at 24 fps, from 5 to 20 seconds per clip.

What aspect ratios are available? Auto, 21:9, 2:1, 16:9, 4:3, 1:1, 3:4, and 9:16.

How do I iterate quickly? Enable draft mode for a fast preview, then disable it for the final render.

What is draft enhance? A draft response returns a cache reference that flux-3-draft-enhance uses to re-render the exact shot at full quality.