FLUX 3 Max release: speed, price, vs FLUX 3
fal released FLUX 3 Max on 19 October 2026: FLUX 3 post-trained for speed, a 10 s clip in about 11 s, #2 on image to video. On Versely from 68 credits.
FLUX 3 Max is a faster, post-trained version of Black Forest Labs' FLUX 3 video model, made by fal Research with BFL and announced by fal on 19 October 2026. It makes 5 to 20 second clips with generated sound at 720p or 1080p, and fal says a 10-second clip takes about 11 seconds. On Versely it starts at 68 credits.
Key facts
| Item | Detail |
|---|---|
| Announced | 19 October 2026, by fal |
| Made by | fal Research, with Black Forest Labs |
| Based on | FLUX 3 Video (BFL, July 2026) |
| Modes | Text to video, image to video, keyframes |
| Clip length | 5 to 20 s, in one-second steps |
| Resolution | 720p; 1080p via upscaling |
| Audio | Generated with the picture, can be off |
| Ranking | #2 on Artificial Analysis image to video |
| On Versely | Since 20 October 2026, 68 to 464 credits |
Sources, read on 27 October 2026: fal's FLUX 3 Max text-to-video and image-to-video pages for the specs and speed claims, Black Forest Labs' FLUX 3 Video page for BFL's own figures, and the Artificial Analysis image-to-video leaderboard for the ranking and release date. fal's announcement on X was summarised by Digg. The speed and quality claims are fal's and BFL's, not Versely's own tests.
What FLUX 3 Max is
Black Forest Labs published FLUX 3 on 23 July 2026 as one model trained across image, video and audio, with clips of up to 20 seconds and sound generated alongside the picture. FLUX 3 Max is not a new foundation model. fal describes it as "a post-trained variant of FLUX 3 Video, tuned for stronger prompt adherence and better visual quality while optimized for lower latency". BFL links to it from its own FLUX 3 Video page with the line "Try FLUX 3 Max on fal".
fal offers three endpoints:
- Text to video: a prompt describes the subjects, action, camera, light and sound. No image needed.
- Image to video: a starting image sets the first frame, the prompt directs the motion, and an optional ending image guides the last frame.
- Keyframes to video: image guidance at several points inside one clip.
Speed: the headline claim
Speed is the reason FLUX 3 Max exists. fal's model pages say it "generates a 10-second video in approximately 11 seconds", against 277.8 seconds for Seedance 2.5 in the same test, which fal calls "more than 25× faster". BFL's page rounds it differently: "10-second video in 12 seconds. 22× faster than Seedance 2.5 at the same quality." fal's post on X also gave about 12 seconds, for a 720p or 768p clip.
fal itself adds the caveat that matters. The comparison "describes the tested benchmark configuration, rather than a guaranteed completion time for every request." fal also says higher reasoning effort increases generation time, and that real times vary with settings and load. On fal's own figures, FLUX 3 Max renders a short clip in roughly the time it plays, which is unusual for a model near the top of the quality tables.
Where it ranks
On 27 October 2026, the Artificial Analysis image-to-video leaderboard listed FLUX 3 Max (720p) at #2 with an Elo of 1,138, from 5,054 appearances. The top five were:
- MiniMax H3 (768p), 1,140
- FLUX 3 Max (720p), 1,138
- Dreamina Seedance 2.5, 1,138
- Gemini Omni Flash 1.1, 1,132
- Wan 3.0, 1,127
The leaderboard gives FLUX 3 Max a possible rank of #1 to #5, which is why fal calls it "joint #1". A two-point gap at this level is inside the margin, so treat the top three as tied. The rating is for the 720p version and for image to video only: fal says its #1 claim "describes the image-to-video evaluation, rather than a separate text-to-video ranking". FLUX 3 Max was not among the top five on the text-to-video boards that day. Our MiniMax H3 vs Gemini Omni post explains how to read these arena gaps.
What is new versus FLUX 3 Video
| Spec | FLUX 3 Video | FLUX 3 Max |
|---|---|---|
| Made by | Black Forest Labs | fal Research, with BFL |
| Focus | Full feature set | Speed and prompt adherence |
| Resolution | Up to 4K (BFL) | 720p, 1080p upscaled |
| Other tools | Editing, continuation, keyframes | Keyframes, reasoning level |
BFL lists editing an existing clip, continuing a video, multiple scenes and up to 4K among FLUX 3 Video's capabilities. fal's FLUX 3 Max pages list none of those. What Max adds is speed, fal's tuning for closer prompt following and cleaner frames, and a reasoning setting (low or high, high by default) that trades time for quality. On fal's API the regular price is $0.17 a second at 720p and $0.29 at 1080p, with a 75% launch discount that runs from 15 October to 7 November 2026.
Who it is for, and who should wait
Good fit:
- Image-to-video work where the first frame is fixed: product shots, portraits, illustrations and stills you have already approved.
- Fast iteration. When a 10-second draft comes back in seconds, you can try five camera moves in the time one slow render takes.
- Ads and social clips that need sound in the file, with framing from 21:9 down to 9:16.
- Shots that must end on a set frame, such as a pack shot, using the optional ending image.
Test first, or wait, if:
- You need native 1080p or 4K detail. fal generates at 720p and upscales to 1080p, and the ranking above is for the 720p output.
- You need keyframes inside the clip. fal's keyframes endpoint is not on Versely yet; the Versely rows take a prompt, or a start image with an optional end frame.
- Budget matters more than speed. On Versely, base FLUX 3 text to video starts at 34 credits, half of Max's 68.
- You mostly make text-to-video clips. Max's ranking is for image to video; check the text-to-video boards for the job you have.
How to use it on Versely
FLUX 3 Max has been on Versely since 20 October 2026, in two modes: text to video, and image to video with an optional last frame. Both are listed on the FLUX 3 model page next to the other FLUX 3 variants. Clips run 5 to 20 seconds in one-second steps, at 720p or 1080p. Text to video offers 21:9, 2:1, 16:9, 4:3, 1:1, 3:4 and 9:16, and image to video adds "auto", which follows the shape of your image.
On 27 October 2026 the model page showed FLUX 3 Max from 68 credits for the shortest 720p clip, and up to 464 credits for 20 seconds at 1080p. Price scales with length and resolution, and the model page always shows today's figure. The base FLUX 3 text to video row runs 34 to 136 credits on the same day.
- Open the FLUX 3 page and pick FLUX 3 Max, text to video or image to video.
- Write a clear order of visible actions: who moves, how the camera moves, the light, and what the scene sounds like. fal's own example ends with "Soft wind, distant gulls and the bicycle's freewheel are audible."
- For image to video, describe what changes after the first frame, not what is already in the picture.
- Add an end frame when the last shot matters.
- Draft at 720p and a short length, then render the keeper at full length.
You can sign up for free and use it on the web or in the iPhone app with one account. For what else moved this month, see our October state of AI video letter.
FAQ
When was FLUX 3 Max released?
fal announced FLUX 3 Max on 19 October 2026, and Artificial Analysis lists the same release date. fal's launch discount is dated from 15 October. It went live on Versely on 20 October 2026.
Is FLUX 3 Max made by Black Forest Labs?
It is built on Black Forest Labs' FLUX 3 Video and post-trained by fal Research in collaboration with BFL. fal serves it, and BFL's FLUX 3 Video page links to it.
How fast is FLUX 3 Max?
fal says a 10-second clip takes about 11 seconds in its benchmark, more than 25 times faster than Seedance 2.5 in the same test. BFL quotes 12 seconds and 22 times. Real times vary with resolution, settings and load.
How much does FLUX 3 Max cost on Versely?
On 27 October 2026 a clip cost 68 to 464 credits, depending on length and resolution. Check the model page for the current price.
What is the difference between FLUX 3 Max and FLUX 3?
FLUX 3 Max is a faster, post-trained FLUX 3 Video tuned for prompt adherence and visual quality, at 720p with 1080p upscaling. BFL's FLUX 3 Video lists more tools, including editing, continuation and up to 4K.
Versely AI Weekly
New AI models, every Thursday.
The image, video, audio and LLM models released that week, what each is good for, and what's new on Versely.