FLUX 3 Max
Explore FLUX 3 Max video inputs, keyframes, duration, resolution and audio controls.
2026.10.20
Make “FLUX 3 Max” exactly the way you wantCreate your own “FLUX 3 Max”
Just ask in plain language — Carat generates optimized images and videos for you.
Start now with 100 free credits daily
No prompts available
What is FLUX 3 Max?
FLUX 3 Max is a video model post-trained by fal Research in collaboration with Black Forest Labs, building on FLUX 3. fal describes improvements in prompt adherence, visual quality and generation latency. The capabilities below describe fal's API. Available options in Carat should be checked in the model picker after release.
Start with words or an image
Text-to-video creates a scene from a written description. Image-to-video begins with a supplied image and can also use an ending image. Describe subject movement and camera movement separately: a slow push-in while a person raises their head gives the model a clearer sequence than a list of style words.
Guide the sequence with keyframes
The keyframes endpoint accepts up to ten images at distinct positions on a 24 fps timeline. This offers guidance for moments within a shot as well as its beginning and end. Choose image-to-video when an opening composition is enough, or keyframes when intermediate compositions matter.
Duration, resolution and sound
The fal API supports 5–20 seconds in one-second steps. Output defaults to 720p, with optional 1080p upscaling. Audio generation can be enabled or disabled, and framing includes portrait, landscape and square formats. Add a speaker and exact dialogue for speech, or describe the environmental sounds you want. Actual waiting time varies with settings and service conditions.
