FLUX 3 Max: fal and Black Forest Labs' video model renders a 10-second clip in about 11 seconds
FLUX 3 Max is a new video model from fal Research and Black Forest Labs that makes 5 to 20 second clips with native sound from text, an image or keyframes. Its makers say a 10-second clip renders in about 11 to 12 seconds, matching Seedance 2.5's score in a preliminary blind test while being far faster.
What it does
Text-to-video, image-to-video and keyframes-to-video in one model, at 720p or 1080p, with audio generated in the same pass. It runs on fal's API (endpoint blackforestlabs/flux-3-max) and is not available as open weights.
Results and limits
In the preliminary Artificial Analysis image-to-video arena it scored 1,139 Elo, the same as Seedance 2.5, but a 10-second clip took about 10.95 to 12.45 seconds against 277.80 seconds for Seedance 2.5. fal says that is 50 Elo points and 11.5 times faster inference than FLUX 3 (126.44 s down to 10.95 s). These are the makers' own and preliminary numbers.
Worth knowing
List price is $0.17 per second at 720p, and launch pricing is 75% off ($0.0425 per second) until November 7. fal published a prompting guide with 50 examples the same day; ComfyUI added FLUX 3 Max partner nodes in v0.42.2.
Try it yourself
Slow dolly-in on [subject]. First [main action], then [second beat]. In the background, [secondary motion]. The light shifts from [time A] to [time B]. Sound: [ambience and one key sound].