FLUX 3 Max - fal’s FLUX 3 Max: 10-second AI video in about 12 seconds

FLUX 3 Max is fal’s post-trained text-to-video and image-to-video model, built with Black Forest Labs on the FLUX 3 base model.

Problem: AI video usually asks you to choose. Waiting minutes for a clip slows down iteration, while faster models often give up fidelity, motion coherence or prompt adherence.

Solution: fal Research post-trained FLUX 3 for video, with a focus on visual fidelity, motion coherence, prompt adherence and image-to-video quality. Per fal, a 10-second 720p, 16:9 clip takes about 12 seconds to generate.

What makes it different: Speed without a big quality drop. fal says it’s about 10x faster than FLUX 3 (12.45 s vs 126.44 s) and about 22x faster than Seedance 2.5 (277.80 s). On Artificial Analysis’s preliminary image-to-video board, it scores 1,139 Elo, tied for #1 with Seedance 2.5, though MiniMax H3 (768p) sits at 1,141 and the board is based on blind human votes. fal highlights faces in close-up, fast-moving subjects and detail retention.

Key features:

• One post-trained model for both text-to-video and image-to-video

• A 10-second clip in about 12 seconds

• 720p output

• Pay per second of video, with no minimums or subscription

• Playground on fal.ai, plus an API with JavaScript and Python examples

• Content generated through the fal.ai API can be used commercially, subject to fal’s terms

Pricing: $0.0425 per second at 720p during the launch promotion, down from the $0.17 list price. The discount ends November 7.

Who it’s for: Creators, studios and developers building video features who want fast iteration without giving up quality.

Try it in the playground or through the API docs. There’s also image-to-video and the FLUX 3 Max page.