Google Veo 3.1

Veo 3.1 turns a written scene into cinematic video with dialogue and sound effects already in sync. Give it a start and end image and it fills in the motion between them, or extend a clip that cuts off too soon.

Optional
Optional
20

Veo 3.1 Key Features

Ingredients to Video

Reference images pin down the things prose is bad at: a specific face, the finish on a specific product, the exact weight of light in a room. Veo 3.1 takes up to three of them and holds what's in them steady through the clip while the prompt drives the action. Describing a color grade in words takes a sentence you'll rewrite four times; handing over a still takes one upload.

Input image 1

Input image 1

Input image 2

Input image 2

Input image 3

Input image 3

Prompt:

Close up shot of woman with sunglasses on top of her head, gold hood earrings, is walking in the interior, she is lost and asks where everyone is and what's going on.

Scene Extension

A single generation runs eight seconds. When the beat you needed lands at nine, the old answer was to generate the whole thing again and hope the second attempt matched. Veo 3.1 takes the last frame as its starting point and keeps going—camera still moving the way it was moving, light unchanged, audio not restarting from silence. Chain a few and you're well past the eight-second ceiling, with nothing to trim at the joins.

Input Video

Prompt 1: Graceful dancer is slowly dancing to classical music.

Prompt 2: A male dancer comes in, gracefully dancing with the woman as classical music plays.

Prompt 3: More dancers show up on the stage.

Prompt 4: The classical music continues, and the dancers continue to dance.

Output Video

Frames to Video

You already know what the opening looks like and what the closing looks like. Veo 3.1 does the part in between—the camera move, the timing, the audio underneath—and you can name the move in the prompt if you want a specific one rather than whatever the model would have chosen. Because both ends are locked to images you picked, the result drops into a timeline without trimming.

First frame

First frame

Last frame

Last frame

Prompt:

Camera soars through the barn to the cowboy riding away.

Output Video

The Same Character, Clip After Clip

Bring the same person back next week, in a different setting, from a prompt that has nothing in common with the first one. Feed Veo 3.1 the reference images you used before—or a frame lifted out of a clip you already made—and the face, hair, and wardrobe come back with them. Everything else is free to change: new location, new light, new camera position, new action. Hold identity across eight or ten generations and you have a cast, which is when short clips start adding up to something.

Input image

Input image

Prompt: a monster walking toward the camera

Output video

Prompt: a cute monster dancing

Output video

Prompt: a cute monster swimming underwater

Output video

Prompt: a cute monster walking in a candy wonderland

Shoot Vertical, Finish in 4K

Veo 3.1 generates 9:16 natively instead of cropping a landscape frame down to it, so a vertical cut keeps its whole composition and nothing important gets shaved off the sides. Resolution goes up to 4K, which matters little for a story post and a lot for anything that lands on a large screen or goes through a second round of grading. Pick the shape and the size before you generate rather than fixing both afterward.

Audio That Comes Out With the Picture

Dialogue, ambience, and effects are generated alongside the frames rather than laid over them afterward, which is why a door closing sounds like it closes when it closes. Veo 3.1 extended this to the reference-image, first-and-last-frame, and extension paths too, so a clip built from stills arrives with sound instead of silent. Speech in very short fragments is still the rough edge—Google says so as well—but ambient and effect work is usually publishable as-is.

Compare Veo 3.1 video models

FeatureVeo 3.1Veo 3.1 FastVeo 3.1 Lite
Best forYour highest-quality, production-ready videosFaster generation at a lower costLowest-cost drafts and high-volume output
Input modesText, image & reference imagesText, image & reference imagesText & image
Reference imagesYesYesNo
Native audioBuilt in — dialogue, effects & ambienceBuilt in — dialogue, effects & ambienceBuilt in — dialogue, effects & ambience
OutputUp to 4K, 4–8s clipsUp to 4K, 4–8s clipsUp to 4K, 4–8s clips
Visual qualityHighestHighGood
SpeedStandardFasterFastest
Relative costStandardLowerLowest
Best choice whenYou need the most polished, client-ready resultYou want speed and savings on everyday clipsYou're iterating fast or producing at volume

GETTING STARTED

How to Generate Video with Veo 3.1

1
Pick Veo 3.1

Select Veo 3.1, then set aspect ratio and resolution. Vertical is generated natively, not cropped. Or start from an image with MaxH3 image-to-video

2
Describe the Shot

Name the subject, the action, and the camera move. Add reference images to lock a face, product, or look.

3
Generate and Extend

One pass gives you eight seconds with audio. Extend from the last frame if you need more, then download the MP4.

Ready to generate your first clip?

YouTube Videos on Google Veo 3.1 AI Video Model

Veo 3.1 fully tested

AI Search

Google Veo 3.1 is INSANE - Full Tutorial

Roboverse

The New Google VEO 3.1 Tested and Explained: Create Ultra-Realistic AI Videos Fast!

AI Edge Mastery

Google Flow | Veo 3.1 Tutorial & Real Test Results

AI2Play

Master Google Veo 3.1: Mind-Blowing AI Video Results (Full Guide)

Invideo for Filmmaking

How To Use Google VEO 3.1 | Create Realistic AI Videos Step-by-step (Full Tutorial)

BIG JOHNAB

Ultimate VEO 3.1 Update EXPLAINED: How To Use Google Veo-3 For Beginners

AI Master

How I Use VEO 3.1 To Create Viral Videos In 9 Minutes

Nicky Saunders

X Posts on Google Veo 3.1 AI Video Model

FAQs

Google DeepMind's video generation model, released October 2025. It produces 8-second clips from a prompt, a still, or up to three reference images, with audio generated in the same pass as the picture.

Veo 3 already had native audio. Veo 3.1 added reference images, first-and-last-frame generation, and clip extension—and brought audio to all three, so a video built from stills is no longer silent.

You get free credits to generate Veo 3.1 clips without entering payment details. Free output carries a MaxH3 watermark; a paid plan removes it.

One generation runs 8 seconds. Scene extension continues from the last frame of a finished clip, so chaining generations gets you past that without a visible cut at the join.

Clean lighting, subject separated from the background, and the thing you care about actually visible—if wardrobe matters, a headshot won't carry it. Three angles of one subject beat three unrelated images.

In the prompt, not a settings panel. Name the move—a handheld tracking shot, a crane up and pull back, a locked-off wide—rather than describing a feeling and hoping the camera agrees.

Try Veo 3.1 on MaxH3

Free credits, no card required. Reference images, first-and-last-frame control, and audio that comes out with the picture—all in one place.

Free generations come watermarked. Upgrade to export clean.