For a Veo app, budget for complete clips and discarded takes. The per-second rate alone does not tell you the minimum clip cost.
Veo 3 API pricing on SpicyAPI starts at $0.045 per second for Veo 3.1 Lite at 720p with sound. Fast works out to $0.09 per second and Veo 3.1 to $0.36. An 8-second clip costs $0.36, $0.72 or $2.88 respectively. Gemini Omni 1.1 Flash, Google's recommended replacement in the Gemini API, costs $0.7297 for 8 seconds at 720p here.
You pay per generation from a prepaid balance, without a subscription or Google Cloud billing account. Choose by task:
- Draft on Lite while changing the scene and camera direction.
- Use Fast for routine Veo work at a middle price.
- Consider Veo 3.1 for the final cut when you want its output and can justify the cost.
- Start a new Google video integration with Omni Flash if you do not depend on a Veo-specific result or extension workflow.
Veo 3 API pricing per second, per 5 seconds and per 8 seconds
All prices are USD for 720p video with generated sound. SpicyAPI effective rates come from per-clip prices; Google's Veo column uses its video-with-audio rates.
| Model | Google list rate per second at 720p with sound | SpicyAPI effective rate per second | 5-second cost equivalent | Actual 8-second clip here |
|---|---|---|---|---|
| Veo 3.1 Lite | $0.05 | $0.045 | $0.225 | $0.36 |
| Veo 3.1 Fast | $0.10 | $0.09 | $0.45 | $0.72 |
| Veo 3.1 | $0.40 | $0.36 | $1.80 | $2.88 |
| Gemini Omni 1.1 Flash | About $0.10136 for video output | About $0.0912 | About $0.456 | $0.7297 |
The 5-second column is a comparison, not a selectable clip length. Veo text-to-video and image-to-video here take 4, 6 or 8 seconds. Omni takes 4, 6, 8 or 10. To deliver a 5-second edit, generate 6 seconds and trim it; budget for the full generated clip.
Shorter clips cost:
| Model | 4-second clip | 6-second clip, enough to trim to 5 seconds |
|---|---|---|
| Veo 3.1 Lite | $0.18 | $0.27 |
| Veo 3.1 Fast | $0.36 | $0.54 |
| Veo 3.1 | $1.44 | $2.16 |
| Gemini Omni 1.1 Flash, 720p | $0.3648 | $0.5473 |
The clip prices appear in our Google video launch update. Google's Cloud pricing page lists the Veo rates. Its Gemini API pricing bills Omni by tokens: $17.50 per million video output tokens and 5,792 output tokens per second of 720p video. Multiplying those gives $0.10136 per output second. Google also bills input tokens, so that output calculation is not the entire direct API invoice.
For Omni on SpicyAPI, use the displayed clip price rather than a rounded per-second estimate. These budgets cover 720p only.
Which tier should you pay for?
Lite: choose framing before spending on a final render
Veo 3.1 Lite is the starting point while the script or motion is changing. At $0.18 for 4 seconds, ten short drafts cost $1.80.
Start short to settle the framing before buying a longer take. Moving to another tier generates a new take, so composition and timing can change.
Fast: the middle option for regular Veo production
Veo 3.1 Fast costs twice as much as Lite and a quarter as much as Veo 3.1 at the same duration: a middle option for regular Veo production.
For shots with several reference pictures, budget for the full reference-to-video task: reference mode on Fast and Lite returns 8 seconds, even for a short insert.
Veo 3.1: reserve the larger budget for a chosen scene
Veo 3.1 costs $1.44 for 4 seconds and $2.88 for 8. Reserve it for a chosen scene rather than repeatedly rewriting a first idea.
We have not run a controlled quality comparison across these four models, so the higher price is no guarantee of fewer retakes. Judge the delivered take before expanding a batch.
Omni Flash: the starting point for a new integration
Gemini Omni 1.1 Flash costs nearly the same as Fast at 720p: $0.3648 versus $0.36 for 4 seconds. Its available lengths extend to 10 seconds, and its reference mode accepts up to seven images rather than the three available on Fast and Lite.
Choose Omni for longer single shots, more references or migration from the retired Gemini API Veo previews. Google names it as their replacement.
Is the Veo 3.1 API still available?
Google's deprecation table lists October 22, 2026 as the shutdown date for the Veo 3.1, Fast and Lite preview models in the Gemini API, with Gemini Omni 1.1 Flash as their recommended replacement.
The shutdown applies to those Gemini API endpoints. Google Cloud still lists Veo pricing, and SpicyAPI lists the three Veo tiers separately. Migrate retired Gemini API model IDs rather than retrying them, and keep model selection configurable in a new app.
Budget for attempts, not just the video you keep
Suppose you need one 8-second final shot. Ten 4-second Lite drafts cost $1.80, and one 8-second Veo 3.1 take costs $2.88: $4.68 in generation charges. If you regenerate the final take twice more, the total becomes $10.44.
Ten 8-second Fast attempts cost $7.20. These examples show how attempts and duration affect the bill; they are not measured success rates. See per-generation costs versus subscriptions for the broader budgeting decision.
All three Veo tiers here output 720p with sound and no silent discount. Extensions charge for the new segment, not the source clip. They continue from the source's last frame and join the segments, without guaranteeing a seamless long take.
Budget for completed clips you discard, too. Most models do not charge for a refused task, and the model page says so when one does. The four Google model pages discussed here state that failed or declined requests are not charged.
Do you need a Google Cloud account or API key?
You need a SpicyAPI account, prepaid balance and API key. No Google Cloud billing account, Google API key or subscription is required.
Choose text-to-video for a written scene or image-to-video for a fixed opening picture. Public input names include duration_seconds, aspect_ratio, and, for an opening frame, image_url. Omni also takes resolution; set it to 720p to match these costs. Do not send a 5-second duration or copy parameter names from Google's SDK into a SpicyAPI request.
Use the model page's request example and the task API documentation to submit and retrieve results. The polling versus webhook guide covers completion handling. Resolve a lost submission response before creating another paid attempt.
Google's content rules still apply. SpicyAPI adds no content filter of its own, but using its API does not remove Google's screening. Our Veo 3 NSFW guide explains the observed limits and refusals.
FAQ
How much does the Veo 3 API cost per video?
At 720p with sound on SpicyAPI, an 8-second Veo 3.1 Lite clip costs $0.36, Fast costs $0.72 and Veo 3.1 costs $2.88. Omni Flash costs $0.7297 for the same length and resolution.
Can I generate a 5-second Veo clip?
Not directly with these tasks. Choose 6 seconds and trim the result to 5. That costs $0.27 on Lite, $0.54 on Fast or $2.16 on Veo 3.1; the 5-second equivalents in the comparison table are arithmetic only.
Is Veo 3.1 Lite the cheapest choice here?
Yes, among these four Google video options at 720p. It starts at $0.18 for 4 seconds. Use it for drafts; choose the final tier based on the result and controls you need.
Does Veo API pricing include audio?
These SpicyAPI prices include generated sound. There is no switch to remove it for a lower price. When comparing Google's own prices, use its video-with-audio rate.
Should a new app use Veo 3.1 or Gemini Omni Flash?
Start with Omni Flash unless you have a reason to keep Veo. Google names Omni as the Gemini API replacement, and it offers longer single clips and more reference slots here. Keep Veo when its output or continuation workflow is the feature you need.
