AI Models
What each model is good at, where it differs from the others, and what it costs to run.
Black Forest Labs' video model. A clip with sound in the same pass — not the Flux 2 / Flux 2 Pro stills rows.
Credits come with a Generator plan. The exact cost is shown on the Generate button before anything is spent.
Flux 3 is Black Forest Labs' video model — the first Flux that moves. It is not a mode of Flux 2 or Flux 2 Pro. Those rows make stills. Flux 3 makes a clip, with sound in the same pass when audio is on. On Visxy it is a video-library shot, or a still you are ready to move.
A take, not a GIF. BFL's pitch is a continuous shot with native audio. The composer maps text-to-video and image-to-video only — not keyframes, not extend, not the /draft ladder. One prompt, or one still plus a prompt.
9:16 matches the video library. Pick a wide frame only when the brief is actually wide.
The Flux comparison on Visxy is stills versus video. Flux 2 and Flux 2 Pro are the photographs. Flux 3 is the moving picture from the same lab. You do not switch a stills row "into video." Animate a Flux still by attaching it here.
Image-to-video takes one picture. There is no last-frame lock on this mapping.
One continuous shot. Paste the library prompt; delete the second city. Use length on purpose — empty seconds drift. Room tone is the honest use of generated audio; replace a line that has to be right.
No Visxy clip is published yet.
No public Flux 3 Edit page. Attach one picture. Describe only the motion. Crop first if the source is the wrong shape.
No. Flux 2 and Flux 2 Pro are image. Flux 3 is video. Same lab, different pages.
Not on Visxy. One reference only.
Attach it to Flux 3. There is no separate Edit article.
