Short answer: Google’s current public documentation identifies production Veo 3.1 models and their access options, but the official material reviewed for this article does not provide a verifiable, current per-second price to quote. It lists fixed quota and Provisioned Throughput as supported consumption options and says pay-as-you-go is not supported for the documented Veo 3.1 models on Gemini Enterprise Agent Platform. So don’t use an old per-second figure as a current Google production API price; confirm the applicable price and quota with Google Cloud for your project before budgeting. (docs.cloud.google.com)
This page covers Google Veo AI video generation, not Veo camera hardware. Pricing and access depend on the API route and model; consumer subscriptions and third-party services are separate offers.
Veo pricing at a glance
| Production API detail | What current Google documentation says |
|---|---|
| Platform | Gemini Enterprise Agent Platform, formerly associated with Vertex AI |
| Veo 3.1 model IDs | veo-3.1-generate-001, veo-3.1-fast-generate-001, and veo-3.1-lite-generate-001 |
| Listed consumption options | Fixed quota and Provisioned Throughput supported; pay-as-you-go not supported |
| Public per-second rate | Not established by the official material reviewed here |
| Gemini API preview status | Veo 3.1 preview model IDs were shut down on October 22, 2026; don’t assume older Gemini API rates or preview IDs still apply |
The standard and Fast models are listed as generally available; Lite is listed as preview. The model documentation also lists U.S. availability in us-central1. These details can affect whether a model fits a production deployment, but they do not establish its price. (docs.cloud.google.com)
Why familiar Veo prices may mislead
You may see articles quoting rates such as $0.05 per second for Veo 3.1 Lite, $0.10 per second for Fast, or $0.40 per second for standard. Those figures appear in secondary pricing coverage, but they should not be presented as current Google production API rates without confirmation from the applicable official price information. The model generation, API route, settings, and effective date matter. (costgoat.com)
Likewise, Google’s September 2025 announcement gave rates for Veo 3 and Veo 3 Fast in the Gemini API. That was a dated announcement for those models and that API surface—not a current rate card for the Veo 3.1 models listed on the production platform today. Google’s Gemini API documentation says its Veo 3.1 preview model IDs were shut down on October 22, 2026. (developers.googleblog.com)
Third-party offers are separate, too. For example, Segmind’s captured Veo 3.1 Fast table lists prices per generation for selected durations and audio settings. Those are Segmind’s prices, not Google API rates, and they cannot be compared directly with a per-second figure.
How to scope a production budget
Because the current official material reviewed here does not establish a public pay-as-you-go rate, there is no defensible Google per-second calculation to publish from it. Instead, ask Google Cloud or your account team for the price and billing basis applicable to your project and model ID, then calculate from that quote.
For a quote billed by generated seconds, a basic estimate would be:
Estimated generation charge = total billable output seconds × quoted rate per second
For example, a plan to create 100 clips at 8 seconds each represents 800 seconds of requested output. If your project’s applicable quote is per second, multiply 800 by that quoted rate. This is a workload example, not a Google price estimate; it assumes each clip is generated once and excludes any separately billed items. If Google quotes fixed quota or provisioned capacity instead, use the quote’s actual unit and terms rather than applying this formula.
Before committing budget, confirm:
- The exact model ID and whether it is GA or preview.
- The API route and consumption option available to your project.
- The quoted currency, billing unit, quota or capacity, and any applicable commitment.
- Whether resolution, duration, or other output settings change the price.
- How retries, failed generations, or outputs beyond the requested amount are treated.
- Whether storage, delivery, or post-processing costs are separate for your setup.
The official model documentation lists Veo 3.1 outputs of 4, 6, or 8 seconds, with supported resolutions and features varying by model. For example, the standard model lists output up to 4K, while Fast lists up to 1080p; both list text-to-video, image-to-video, and sound generation capabilities. Check the exact model page for the settings your workflow needs. (docs.cloud.google.com)
Which Veo model should a production team assess?
- Veo 3.1 standard (
veo-3.1-generate-001): Consider assessing it when the workflow needs the model’s listed 4K output option or first-and-last-frame generation. It is listed as GA. (docs.cloud.google.com) - Veo 3.1 Fast (
veo-3.1-fast-generate-001): Consider it for workflows that value faster generation; Google describes Fast as designed for speed and standard production workflows. It is listed as GA, with output up to 1080p in the model documentation. (cloud.google.com) - Veo 3.1 Lite (
veo-3.1-lite-generate-001): Google describes Lite as its cost-effective tier for high-volume applications, but the current model documentation lists it as preview. Verify preview terms and suitability before making it a production dependency. (cloud.google.com)
Google’s model documentation supports model selection by capability, but it does not make these models’ prices interchangeable. Compare them using the price and billing terms for the exact route and model you plan to deploy—not historical articles or another provider’s per-generation table.
Bottom line
For production API budgeting, the reliable answer is not a universal Veo price per second. Current Google documentation lists fixed-quota and Provisioned Throughput options—not pay-as-you-go—for the Veo 3.1 models on Gemini Enterprise Agent Platform, while the official material reviewed here does not establish a current public dollar rate. Obtain project-specific pricing for your selected model and route, then estimate from its actual billing unit. Also treat the former Gemini API Veo 3.1 preview rates and IDs as obsolete for new planning: Google states those previews shut down on October 22, 2026. (docs.cloud.google.com)
Sources
Veo 3.1 | Gemini Enterprise Agent Platform | Google Cloud Documentation
Generate videos with Veo 3.1 in Gemini API | Google AI for Developers
Agent Platform Pricing | Google Cloud
Veo 3.1 Lite and a new Veo upscaling capability on Vertex AI | Google Cloud
Veo 3 and Veo 3 Fast – new pricing, new configurations and better resolution