All models

Google

Veo 3.1

Google's preview video model for text-, image-, and video-guided generation with native audio and output options up to 4K.

Plain-English overview

What Veo 3.1 actually is

Veo 3.1 handles short clips with synchronized audio and supports generation, extension, interpolation, and reference-driven workflows through variant-specific endpoints. Resolution, duration, and input mode are linked; not every combination is available.

Google publishes an unusually useful request-latency range of 11 seconds to six minutes during peak periods. That is an operational range, not a quality benchmark, and higher resolution can increase latency and price.

Good fit for

  • Short cinematic clips with native audio
  • 4K experiments and video extension
  • Google-centric media pipelines

Category comparison

The facts that matter for video models

These are provider-published specifications, not Cody benchmark scores. Follow the linked sources for current limits and endpoint-specific exceptions.

Generation priceCurrent list price per generated second or the provider's closest billing unit.
$0.40/s standard with audio
Clip durationDocumented duration choices or maximum clip length.
4, 6, or 8 seconds
Maximum outputHighest provider-documented output resolution, with relevant constraints.
Up to 4K; 1080p/4K require 8s
Native audioWhether the endpoint generates synchronized audio with video.
Yes
Input modesSupported text, image, video, reference, or motion inputs.
Text, image, or video to video
Generation timeOnly shown when a provider publishes a range; not a Cody benchmark.
Official range: 11s–6m

API and provider access

Where to get Veo 3.1

Availability

Regions and access stage

Preview access through Gemini API; variants are also available through selected platforms.

Google publishes regional person-generation restrictions for EU, UK, Switzerland, and MENA locations.

Check live availability

Data and training

The route matters.

Google says paid Gemini API content is not used to improve its products. Generated videos are retained on the server for two days and should be downloaded during that window.

This is a concise reading of the cited provider material, not legal advice. A third-party gateway can have different storage, routing, training, and residency terms from the model maker's direct API.

Read the provider policy

Frequently asked questions

Veo 3.1 FAQ

What is Veo 3.1?

Google's preview video model for text-, image-, and video-guided generation with native audio and output options up to 4K. Veo 3.1 handles short clips with synchronized audio and supports generation, extension, interpolation, and reference-driven workflows through variant-specific endpoints. Resolution, duration, and input mode are linked; not every combination is available.

When was Veo 3.1 released?

Veo 3.1 was released on September 1, 2025 according to the cited provider materials.

Where can I access Veo 3.1?

Preview access through Gemini API; variants are also available through selected platforms. The access routes listed in this guide are Google, fal, Runway.

How much does Veo 3.1 cost?

$0.40 / second at 720p or 1080p. Standard output with audio is $0.40/s at 720p or 1080p and $0.60/s at 4K. Fast and Lite variants have different rates.

Where is Veo 3.1 available?

Preview access through Gemini API; variants are also available through selected platforms. Google publishes regional person-generation restrictions for EU, UK, Switzerland, and MENA locations.

Is my Veo 3.1 API data used for training?

Google says paid Gemini API content is not used to improve its products. Generated videos are retained on the server for two days and should be downloaded during that window. The policy belongs to the provider route and account terms, so verify it again before production use.