Video model · by Google DeepMind
Veo 3.1 AI Video Generator
Google's most capable video model — dialogue, sound effects, and cinematic camera work in one generation.
Last updated July 11, 2026
Veo 3.1 is Google DeepMind's video generation model, and the first Veo that treats audio as a first-class output: speech, ambient sound, and effects are generated together with the picture, in sync. On AI Canvas you can run it from a text prompt or animate an existing image, choose the clip length and resolution before you spend credits, and keep every result in your library.
Example prompts for Veo 3.1
Close-up of a fisherman in a yellow raincoat on a stormy pier, he looks at the camera and says "the sea keeps what it takes", waves crashing behind him, cinematic lighting
Slow orbital shot of a matte black wristwatch on a marble pedestal, soft studio light sweeping across the dial, shallow depth of field, no text
Aerial shot descending through fog over a pine forest at dawn, a river cutting through the valley, birdsong and wind, documentary style
What Veo 3.1 is good at
Native audio, not a music bed
Veo 3.1 generates dialogue, ambient noise, and sound effects as part of the video itself. Characters speak the lines you write in the prompt, and footsteps land when feet touch the ground — no separate audio pass needed.
Text to video and image to video
Start from a written scene or upload a frame and direct what happens next. Image to video keeps the subject, style, and framing of your still while adding motion you describe.
Real camera language
The model understands shot vocabulary — orbit, dolly, crane, rack focus, handheld. Prompts written like shot lists come back looking like coverage, not screensavers.
How to use Veo 3.1 on AI Canvas
- 1
Describe the shot or drop in a frame
Write the scene the way you'd brief a camera operator — subject, action, camera move, mood. Or switch to image to video and upload the frame you want to animate.
- 2
Pick length, resolution, and aspect ratio
Choose the clip duration and output quality, and set the aspect ratio for where the video will live — widescreen, vertical, or square. The credit cost updates before you generate.
- 3
Generate and iterate
Results land in your library in a few minutes. Rerun with a tweaked prompt, extend the shot, or send the clip to the video tools for captions, upscaling, or reframing.
Who it's for
For short-form creators
Vertical clips with generated speech and sound are ready for TikTok, Reels, and Shorts without an editing pass. Hook lines can be spoken by the character on screen.
For marketing teams
Product-in-scene shots, seasonal spots, and concept ads in hours instead of shoot days. Iterate on the script by editing a sentence, not rebooking a studio.
For filmmakers and storyboarders
Previsualize scenes with real camera moves and temp dialogue before committing to a shoot. Image to video turns concept art into moving boards.
Other models to try
Sora 2 Pro
OpenAI's most advanced video model — startlingly real motion, physics, and sound, up to 12 seconds per clip.
Kling v3 Pro
Kling's flagship model — multi-shot scenes, native audio, and the most cinematic motion in its class.
Nano Banana Pro
Google's viral image model, pro tier — sharper detail, better text, and edits that keep faces intact.
Compare & guides
Frequently asked questions
Does Veo 3.1 generate sound and dialogue?
Yes. Veo 3.1 generates audio natively — spoken dialogue you write in the prompt, ambient sound, and effects, all synchronized with the picture. You can also turn audio generation off to save credits.
Can I use Veo 3.1 for free?
AI Canvas is credit-based with no subscription required — new accounts get free credits to try models like Veo 3.1, and you only buy credits when you need more.
What is the maximum resolution and clip length?
Veo 3.1 on AI Canvas generates clips of 4, 6, or 8 seconds at up to 4K resolution, in 16:9, 9:16, or 1:1 aspect ratios.
What's the difference between Veo 3.1, Veo 3.1 Fast, and Veo 3.1 Lite?
Same family, different trade-offs: Veo 3.1 is the highest quality, Fast cuts generation time and cost with a small quality trade-off, and Lite is the cheapest way to get Veo output with audio. All three are available on AI Canvas.
Veo 3.1 or Sora 2 — which should I use?
Both generate video with audio. Veo 3.1 tends to follow camera direction more literally and offers 4K output; Sora 2 excels at physical realism and longer 12-second clips. On AI Canvas you can run the same prompt through both and compare.
Can I use the videos commercially?
Yes — videos you generate on AI Canvas are yours to use, including in commercial projects like ads, client work, and monetized content.
Make your first video with Veo 3.1
Free credits on signup. No subscription — pay only for what you generate.
Start Creating