Google Veo AI Video: 3.1 and 3.1 Fast

Veo versions

Version	Status	Notes
Veo 3.1video	Live	Improved consistency and scene extensions.	Try now →

Why creators cast Veo

01

Sound is part of the shot

Veo generates sound effects, ambient noise and dialogue natively. No post pass on short work.

02

Cinematic by default

Google DeepMind re-designed Veo for greater realism. Light, lens and physics read like footage.

03

Two lanes, one wallet

Veo 3.1 renders 720p from ⚡326 and 1080p from ⚡652. Veo 3.1 Fast drafts from ⚡123.

Credits, not seats

Pay per render. See the price first.

Veo 3.1

From ⚡326. 720p ⚡326 for 4s. 1080p ⚡652 for 8s.

Veo 3.1 Fast

From ⚡123. 720p ⚡123 for 4s. 1080p ⚡245 for 8s.

Video is paid per render, always. Veo 3.1 starts at ⚡326 per clip. Veo 3.1 Fast starts at ⚡123. A Veo render costs more than the ⚡50 signup grant, so treat the grant as a first look at the composer. $10/1,000⚡ The composer quotes the exact ⚡ before anything renders, and you approve it.

Frequently asked questions

Veo is free to try on OpenClips, not free to run at volume. Make a free account and ⚡50 lands in your wallet with no card. Video is billed per render: Veo 3.1 starts at ⚡326 and Veo 3.1 Fast starts at ⚡123. A Veo take costs more than the signup grant, so top up when you're ready. The composer quotes the exact ⚡ before you render.

Veo is Google DeepMind's video model family, built for cinematic shots with sound baked in. Google DeepMind describes Veo under the tagline "Video, meet audio." On OpenClips you reach it through the same composer as every other model, on one credit wallet and one library. No separate vendor subscription, no export dance between tools, no waiting list.

Veo clips run 4, 6 or 8 seconds per render. You pick the length in the composer and the price follows it: ⚡326 at 4 seconds, ⚡489 at 6 seconds, ⚡652 at 8 seconds. 1080p renders at 8 seconds only, so that clip is ⚡652. Need more screen time? Render a few takes and cut them together.

Open the composer, pick Veo, write your prompt, then render. Describe the camera move, the action and the audio in the same prompt. Improved prompt adherence means more accurate responses to your instructions. Add up to 3 reference images to lock a face, a product or a location. Choose 16:9 or 9:16, check the ⚡ quote, then render.

Yes, Veo writes the sound with the picture. Google DeepMind says Veo offers new levels of control, consistency, and creativity, now across audio. Prompt the sound the way you prompt the shot: rain on glass, muffled traffic, one whispered line. It comes back in sync, so short pieces skip the sound design pass entirely.

Cast Veo when the shot needs dialogue, ambience and a cinematic look. Cast Seedance when you want longer takes and heavier reference control. You don't have to guess between them: both live in the same composer on one wallet, so run the prompt twice and keep the better take. Drafting on Veo 3.1 Fast at ⚡123 keeps the test cheap.

Veo can be used in Gemini and in Google Flow, and developers can build with it through the Gemini API video docs. OpenClips is the consumer route to it: one login, one credit wallet, and every other video model sitting beside it. No stack of per-vendor plans. Start free, pay per render, keep every clip in one library.

Your next shot has a soundtrack.

Free account. Every model in one composer. Pay per render.

Generate with Veo 3.1

⚡50 on signup. No card. Pay per render.

Sound arrives with the picture

Most video models hand you silence. Veo hands you a scene with audio already in it. You write the camera move, the action and the ambience in a single prompt. Veo lets you add sound effects, ambient noise, and even dialogue to your creations, generating all audio natively. That kills the post pass on short work: dialogue beats, product ASMR, rain on a window at night.

Cinematic defaults

Veo's light reads like a lens, not a filter. Golden hour exteriors, rim-lit hero shots, shallow-focus conversations. Google DeepMind says the model delivers high quality, excelling in physics, realism and prompt adherence. That's why it gets cast for ads and film fragments rather than quick memes. Give it camera direction and it follows.

Two lanes, one wallet

Veo 3.1 is the finishing lane, from ⚡326. Veo 3.1 Fast is the draft lane, from ⚡123. Both render 720p and 1080p, in 16:9 or 9:16, at 4, 6 or 8 seconds. Both take up to 3 reference images to hold a face or a product steady. Draft cheap, finish clean, keep every take in the same library.

Where to go next

Open the AI Video Generator to see the whole roster in one composer. Run the same prompt on Veo and on a rival cast, then keep the winner. Come back to Veo when the shot needs a voice.