Content Creators
Turn short ideas into vertical, square or landscape clips, with sound described alongside the action.
Turn text, a starting image or timed keyframes into 5–20-second videos with optional sound. Create with FLUX 3 Max on Picir AI now, with 75% off generation credits before November 7, 2026 at 00:00 UTC.
A video model built on FLUX 3, post-trained by fal Research for detailed motion and faster generation.
FLUX 3 Max is a video generation model from fal Research, released on October 19, 2026 and developed with Black Forest Labs using FLUX 3 as its foundation. Additional training refines how it follows prompts and renders motion, while inference optimizations reduce generation time.
FLUX 3 Max supports three ways to create a video:
Max supports 720p and upscaled 1080p. The original FLUX 3 also offers 1440p and 2160p through its official service, but both models currently offer only 720p and 1080p on Picir AI.
Generated audio is optional. Describe dialogue, music or environmental sound in the prompt, or turn audio generation off when you plan to add your own soundtrack.
Explore examples of coordinated action, close-up detail and imaginative scenes.
FLUX 3 Max
A horseback mail exchange
FLUX 3 Max
A small gesture in close-up
FLUX 3 Max
A driverless car in the Old West
FLUX 3 Max
A robot dog working the herd
FLUX 3 Max
A call from the future
FLUX 3 Max
Muscle from the future
Explore faster iteration, flexible inputs, timed guidance and optional sound for short video creation.
The model's published benchmark reports 12.45 seconds for a 10-second, 16:9 clip in a 720p/768p-class comparison. That offers room for faster iteration, though actual waiting time varies with settings and demand.
Describe a subject, its actions, the camera and the light to create a scene without a starting image. Clear sequences of actions help communicate how the shot should unfold.
Use a still image to establish the opening composition, then describe the movement. An optional ending image guides the final frame for transitions between two visual states.
Position up to 10 images along the video timeline to guide intermediate moments. Timed keyframes offer more control over a sequence than an opening and closing image alone.
Include dialogue, narration, music or environmental sound in the prompt. Audio generation is optional, so you can also create silent clips for your own soundtrack.
Shape a 5–20-second clip for widescreen, square or vertical viewing. The documented options include seven aspect ratios, automatic framing, 720p and upscaled 1080p.
From social clips to visual concepts, these workflows show who can use the model and what they can create.
Turn short ideas into vertical, square or landscape clips, with sound described alongside the action.
Animate an illustration, portrait or concept image while using the original composition as the starting point.
Explore camera moves and short narrative scenes, or use timed images to guide important moments in a sequence.
Develop short visual concepts and variations from a written brief or an existing image before choosing a direction.
A simple workflow for the model's documented generation modes.
Write the subject, action, camera movement and sound. For image-led scenes, prepare a starting image, an optional ending image or a sequence of timed keyframes.
Choose a duration from 5 to 20 seconds, an aspect ratio and 720p or 1080p output. Decide whether the clip should include generated audio.
Preview the generated video and check the motion, composition and sound. Adjust the prompt or guiding images to explore another version.
Model capabilities documented by fal; these describe the model, rather than the controls of other Picir AI generators.
| Developer | fal Research, in collaboration with Black Forest Labs |
| Foundation | FLUX 3 |
| Generation modes | Text to video; image to video; timed keyframes to video |
| Clip duration | 5–20 seconds in one-second increments |
| Output resolution | 720p; upscaled 1080p |
| Aspect ratios | Auto, 21:9, 2:1, 16:9, 4:3, 1:1, 3:4, 9:16 |
| Image guidance | Starting image with optional ending image; up to 10 timed keyframes |
| Keyframe timing | 24 fps timeline; unique positions from frame 0 through duration × 24 |
| Input image formats | PNG, JPEG, WebP |
| Audio | Optional generated audio; enabled by default |
| Output format | MP4 in published examples |
Last updated: October 20, 2026
Compare fal's published model capabilities. Max prioritizes faster generation; the original offers higher resolution tiers through its official service. On Picir AI, both models currently offer 720p and 1080p.
| Specification | FLUX 3 Max | FLUX 3 |
|---|---|---|
| Model approach | Post-trained FLUX 3 variant | Original foundation model |
| Clip duration | 5–20 seconds | 5–20 seconds; automatic duration also available |
| Resolution tiers | 720p, 1080p | 720p, 1080p, 1440p, 2160p |
| Higher resolution processing | 1080p is upscaled | 1080p, 1440p and 2160p use an upsampler |
| Aspect ratios | Auto plus seven ratios | Auto plus seven ratios |
| Generated audio | Optional | Optional |
| Reasoning effort control | Low or high | Not listed in the image-to-video documentation |
FLUX 3 Max Advantages: Prioritizes generation speed while refining motion and prompt adherence. It supports text, a starting image or timed keyframes, with optional generated audio.
FLUX 3 Advantages: Offers additional 1440p and 2160p output tiers. These higher resolutions use an upsampler, and its documented workflow also offers automatic duration.
Choose FLUX 3 Max for fast iteration at 720p or 1080p. Both models offer these two resolutions on Picir AI; the original model's official 1440p and 2160p tiers are not available here. Check the selected model's displayed credit cost before generating.
Choose a Picir AI plan and check FLUX 3 Max generation costs below.
FLUX 3 Max generation is 75% off before 2026-11-07 00:00 UTC. At 720p, the rate drops from 30 to 7.5 credits per second; at 1080p, from 50 to 12.5 credits per second. During the promotion, the table shows the total credits per clip after the 75% discount.
| Resolution / duration | Credits per generation |
|---|---|
| 720p / 5 seconds | 38 |
| 1080p / 5 seconds | 63 |
| 720p / 10 seconds | 75 |
| 1080p / 10 seconds | 125 |
| 720p / 20 seconds | 150 |
| 1080p / 20 seconds | 250 |
Credits are calculated for the full clip and rounded up once to a whole credit. For example, a 5-second 720p clip costs 38 credits during the promotion. Check the final total in the generator before submitting.
The discount applies only to FLUX 3 Max generation credits, not subscription plans. Standard rates resume at the cutoff.
Browse other video models and tools for your next creative project.
Find answers about supported inputs, video length, resolution, audio, generation speed and costs.
FLUX 3 Max is a video generation model built on FLUX 3, with additional training to improve motion and prompt adherence, plus optimizations for faster generation.
The model supports text, a starting image with an optional ending image, and up to 10 timed keyframes. On Picir AI, you can currently generate from text, a single image, or a pair of start and end frames. Timed keyframes are not yet available on this site and will be added later.
The model supports 5–20 seconds per generation in one-second increments. A longer continuous output is not a documented single-generation option.
Its documented resolution choices are 720p and upscaled 1080p. The original FLUX 3 has higher resolution tiers, which should not be confused with Max's options.
Audio generation is optional, and official examples include dialogue, narration, music and environmental sound. The reviewed controls do not include a named voice selector or a reference-audio upload.
The model's published benchmark reports 12.45 seconds for a 10-second, 16:9 clip in a 720p/768p-class comparison. This is a benchmark rather than a guaranteed completion time; settings and service demand affect the wait.
Describe the subject and action first, then add camera movement, lighting and sound. For image-to-video, explain what should move from the supplied image and keep the sequence achievable within the chosen length.
Before 2026-11-07 00:00 UTC, FLUX 3 Max generation credits are 75% off: 720p costs 7.5 instead of 30 credits per second, and 1080p costs 12.5 instead of 50. The full clip cost is rounded up to a whole credit. Check the final total in the generator; subscription plans are not discounted by this offer.
You can use the FLUX 3 Max AI Video Generator on Picir AI.