Veo 3.1 video generator
- Type
- Video model
- Made by
- Google DeepMind
- Input
- Text, plus a starting image, or a first and last frame
- Output
- MP4 video, with or without sound
- Options
- 2 aspect ratios · Duration 4, 6, 8 s · Resolution: 720p, 1080p · Sound · First & last frame
- Price
- From 800 credits
- Plan
- Paid plans
Veo 3.1 is Google DeepMind's video model. It makes high-fidelity clips with synchronized sound, from a text prompt, a starting image, or a first and last frame. Clips are 4, 6 or 8 seconds long, in landscape or vertical, at 720p or 1080p.
What you can do with it
- Text to video. Describe a shot and get a short clip with sound.
- Start from an image. Attach a picture and the first image becomes the opening frame.
- First and last frame. Switch on First & last frame, add the first frame (required) and an optional last frame, and the clip moves between them.
- Silent clips. Switch Sound off when you plan to add your own audio.
How to use it
- Write the shot in the composer.
- Attach a starting image, or switch on First & last frame and add your frames.
- Choose aspect ratio, duration, resolution and sound.
- Check the price in the composer, then generate.
Writing prompts for Veo 3.1
With at most 8 seconds per clip, focus each prompt on one beat.
- Subject and action first. Who or what is on screen and what they do.
- Then setting and look. Place, time of day, light, lens feel, color.
- One camera move. A slow push-in, a handheld follow, a static wide.
- Describe the sound you expect. Footsteps on a wooden floor, a café's background chatter, waves. With Sound on, the audio is generated with the picture.
- Use Avoid for what you don't want. The Avoid field in the advanced settings panel is a place to list things to keep out of the clip, such as on-screen text or extra people.
Example prompts:
- A barista pours steamed milk into a latte, forming a leaf pattern. Tight overhead shot, morning light. The hiss of the steam wand and quiet café chatter.
- A golden retriever bounds through fresh snow toward the camera, kicking up powder. Low angle, slow motion feel. Muffled crunching snow, a distant whistle.
Choosing settings
- Aspect ratio: 16:9 (default) or 9:16.
- Duration: 4, 6 or 8 seconds, default 8.
- Resolution: 720p (default) or 1080p.
- Sound: on by default; switch off for a silent clip.
- First & last frame: off by default.
- Avoid and Seed: optional text fields in the advanced settings panel. Avoid lists what to keep out of the clip; Seed lets you pin the random starting point of a run.
Veo 3.1 costs 200 credits per second of video, and the rate is the same at 720p and 1080p, with sound on or off. A 4-second clip is 800 credits, 6 seconds is 1,200 and 8 seconds is 1,600. Since resolution doesn't change the price, 1080p is a sensible choice for final takes. The composer shows the exact price before you generate.
Need a longer shot? Seedance 2.0 goes up to 15 seconds and Wan 3.0 up to 30, both with a Sound switch.