Fal released FLUX 3 Max, a post-trained text-to-video model. It is a variant of FLUX 3 Video tuned for better prompt adherence and visual quality. It is optimized for lower latency. It ranks joint first in Artificial Analysis image-to-video benchmarks. It generates a 10-second video in about 11 seconds. It supports 5–20 second clips in one-second steps. It outputs 720p by default with optional 1080p upscaling. It can generate audio alongside the video. It offers flexible aspect ratios from 21:9 to 9:16. It provides low or high reasoning levels. Promotional pricing runs from October 15 to November 7. Standard rates are $0.17/second at 720p and $0.29/second at 1080p after the promo. It is available via API and playground on fal.ai.
What changed
FLUX 3 Video had no post-trained variant optimized for latency and prompt adherence.
What it unlocks
Generating 5–20 second videos from text prompts with native audio at 720p or 1080p.
- 10-second video in ~11 seconds
- 25× faster than Seedance 2.5
- $0.0425/second at 720p (promo)
- $0.0725/second at 1080p (promo)
What you need to act on it
- a fal.ai account
- API access or playground use
Sources