Veo 3.1 is Google's video model. On this site it runs through fal, and the numbers on the right are the limits we offer, taken from fal's published Veo 3.1 model card (checked 2026-10-19). Each Veo 3.1 clip is one continuous shot of 4, 6 or 8 seconds at 1080P, and the sound is generated together with the picture.
* Provisional rate for this model; it may change before the model is generally available. The Generate button always shows the exact amount for your settings.
Describe the subject, what it does, where the camera is and what we should hear. The model reads the whole description as a single shot, so a cut or a second location belongs in a second clip. The same flow works on every model in text to video, if you want to compare Veo 3.1 with another model on one prompt.
Upload a still and say how it should move. The frame sets the look and the aspect ratio; the prompt sets the motion, the camera and the sound. Veo 3.1 here takes one opening frame only, not an end frame. More on frames in image to video.
Write the lines out and name who speaks. fal's own Veo 3.1 example does exactly that, quoting both sides of a two-person street interview.
Rain on a window, a crowd, a kettle: sound you describe has a better chance of landing in the Veo 3.1 track than sound you leave to chance.
We request every clip with audio, and the price is set for that. If you need silence, say so in the prompt.
Both have the same limits here: 4, 6 or 8 seconds, 1080P, sound on, text or a first frame. Fast costs fewer credits per second, so many people draft on Veo 3.1 Fast and render the keeper on Veo 3.1. Need 15 seconds or a 2K frame? That is MiniMax H3, not Veo 3.1.
| Model | Inputs | Length | Resolution | Credits per second |
|---|---|---|---|---|
| MiniMax H3 | Text, Image, References | 4–15s | 768P / 2K | 200 at 768P, 300 at 2K |
| Veo 3.1 | Text, Image | 4s / 6s / 8s | 1080P | 1,100 at 1080P* |
| Veo 3.1 Fast | Text, Image | 4s / 6s / 8s | 1080P | 413 at 1080P* |
* Provisional rate; it may change before the model is generally available.
Not sure what to write? Every prompt in the library can be loaded into the console above in one click.
Browse video promptsVeo 3.1 is Google's video generation model. It makes short single-shot clips from a text prompt or from a starting image, and generates sound together with the picture. On this site Veo 3.1 runs through fal.
Veo 3.1 clips here are 4, 6, 8 seconds long, at 1080P, in 16:9 or 9:16. fal also lists 720p and 4K for Veo 3.1; we only offer 1080P for now.
Yes, from one opening frame. Upload the still, then describe how it should move and what it should sound like. Veo 3.1 on this site does not take a closing frame.
It is priced per second of output: 1,100 credits per second at 1080P, so a 4-second Veo 3.1 clip reserves 4,400 credits. That rate is provisional and may change before the model is generally available. The Generate button shows the exact amount before you submit, and failed tasks return their credits.
On this site, yes. Every Veo 3.1 request asks for audio and the price is set for that. If you want a quiet clip, say so in the prompt.