Short answer: The Google documentation retrieved for this update does not provide an exact Veo 3.1 production rate or billing unit, so it is not enough to calculate a verified Google API price. Google’s Agent Platform model page points to a separate official Agent Platform pricing page, but the price details and billing terms needed to quote a Veo 3.1 cost are not established here. Don’t treat third-party per-second or per-video figures as Google’s rates.
Which Veo 3.1 API route is available?
Google says the Veo 3.1 Preview models in the Gemini API—including Generate, Fast Generate, and Lite Generate—were shut down on October 22, 2026. Its migration guidance points developers to Gemini Omni Flash or generally available models on Gemini Enterprise Agent Platform. This shutdown applies to those preview model IDs; it should not be generalized to every Google video-generation product. (ai.google.dev)
Google’s Agent Platform documentation lists these Veo 3.1 model IDs: veo-3.1-generate-001, veo-3.1-fast-generate-001, and veo-3.1-lite-generate-001. It marks Generate and Fast Generate as GA and Lite as Preview on that platform. The same page lists Provisioned Throughput and Fixed quota as supported consumption options, while marking pay-as-you-go unsupported for these models. These are Agent Platform details, not Gemini API billing terms. (docs.cloud.google.com)
What does official pricing establish?
The Veo model page directs readers to the Veo section of Google’s Agent Platform pricing information, but the price and billing unit are not stated in the model documentation itself. The pricing material retrieved for this article does not establish a Veo 3.1 rate that can be quoted here. Check the official pricing page for the applicable product and billing arrangement before approving a budget. (cloud.google.com)
That distinction matters because the model documentation does not support a simple pay-as-you-go assumption for its listed Agent Platform Veo models. Confirm the price and any applicable quota or throughput arrangement for the exact model and project. The documentation also says the Gemini API is not a Google Cloud offering and has its own terms. (docs.cloud.google.com)
Third-party pages publish Veo price estimates, but those are not a substitute for a current Google rate. For example, reported per-video amounts do not by themselves specify the model route, output settings, or official billing basis. Treat them as unverified comparisons, not as Google API prices.
How to estimate a production workload
Start with the official rate and billable unit for the route you plan to use. Then calculate:
Estimated generation charge = confirmed rate × confirmed billable usage
For a workload of 500 clips at 8 seconds each, the requested output totals 4,000 seconds (500 × 8). That is an output-volume calculation—not a bill—unless the applicable Google pricing terms confirm that output seconds are the billable unit and explain how the selected model and settings are charged.
Before turning that volume into a dollar estimate, verify the rate, unit, and any relevant settings or charges for your chosen arrangement. The available model documentation lists 4-, 6-, or 8-second generation lengths and up to four output videos per prompt; those technical limits do not, by themselves, define billing. The supplied pricing evidence also does not establish how retries, failed generations, or multiple outputs are charged. (docs.cloud.google.com)
Which model fits the output requirement?
Google’s Agent Platform documentation lists these capability differences:
| Agent Platform model | Listed status | Output resolutions listed | Reference images |
|---|---|---|---|
Generate (veo-3.1-generate-001) |
GA | 720p, 1080p, 4K | Supported |
Fast Generate (veo-3.1-fast-generate-001) |
GA | 720p, 1080p | Supported |
Lite Generate (veo-3.1-lite-generate-001) |
Preview | 720p, 1080p | Not supported |
These are platform-specific capabilities, not a comparison of price or output quality. The page lists United States availability in us-central1; that listing does not establish availability in other regions or on other routes. If your workflow requires 4K, Generate is the only one of these three models whose listed output resolutions include it. If you need reference images, the page lists Generate and Fast as supporting them, but not Lite. Confirm availability and price for your account and region before selecting a model. (docs.cloud.google.com)
API pricing is not the same as subscription pricing
Google AI or Flow subscriptions and third-party subscriptions are different purchase routes from production API or cloud billing. A monthly plan price, credit allowance, or vendor’s advertised video count is not a Google API rate. Google’s consumer plan page describes video-generation benefits but does not establish a Veo 3.1 API price. (docs.cloud.google.com)
Bottom line: Google’s retrieved documentation identifies Agent Platform Veo 3.1 models and directs readers to official pricing information, but the material available here does not establish an exact production price or billing unit. The Gemini API Veo 3.1 Preview models were shut down on October 22, 2026; evaluate an available route separately rather than assuming those preview rates or IDs still apply. Use the official pricing page to confirm the applicable rate and billing arrangement before turning workload volume into a dollar budget. (cloud.google.com)
Sources
Agent Platform Pricing | Google Cloud
Veo 3.1 | Gemini Enterprise Agent Platform | Google Cloud Documentation
Generate videos with Veo 3.1 in Gemini API | Google AI for Developers