Veo 3 Pricing in 2026 and What Replaced It for AI Video Ads

Veo 3 pricing no longer exists in the Gemini API. Google shut down Veo 3.0 on 30 June 2026 and Veo 3.1 Preview (standard, fast and lite) on 22 October 2026, and points developers to Gemini Omni Flash instead. Read on 30 October 2026, Omni Flash is billed by tokens, which Google converts to about $0.10 a second of 720p video. That makes a 15-second ad roughly $1.52 and a 30-second ad roughly $3.04 per attempt, before retries.

Making ads, not clips? A cheap per-second rate still leaves you writing the script, the hook and the caption timing. See AdsGen plans and paste a product URL to get UGC scripts and 9:16 videos with burned-in captions.

What happened to Veo 3 pricing

Most articles you will find for this phrase quote per-second rates for Veo 3 and Veo 3.1 Fast. Those rates describe models you can no longer call. Google's deprecations page lists veo-3.0-generate-001 with a shutdown date of 30 June 2026 and gemini-omni-1.1-flash as the recommended replacement. The video guide in the Gemini API docs adds that Veo 3.1 Preview, including the fast and lite variants, was shut down on 22 October 2026.

The same notice says teams can also use "the GA models available through the Gemini Enterprise Agent Platform". We did not find a published per-second Veo price for that route, so we do not quote one. If a page gives you an exact Veo 3 rate today, check whether it names a model you can still call and the date it was read.

Gemini Omni Flash pricing

Omni Flash is the model Google now documents for video generation and conversational editing in the Gemini API. It is on the paid tier only. The pricing page lists the standard rates per one million tokens:

A footnote on the page converts that into video terms. Billing is based on total output token use, at 5,792 tokens for each second of 720p video, which Google says equates to approximately $0.10 a second under standard pricing. Our own arithmetic agrees: 5,792 tokens at $17.50 per million is $0.1014 a second. The input side is billed on top, though for a short text prompt it is a fraction of a cent.

What one Veo-era ad costs on Omni Flash

Ad length is what turns a per-second rate into a budget. These figures use the $0.1014 a second we calculated at 720p and ignore input tokens.

The attempts columns matter more than the first one. Almost nobody keeps the first generation of an ad. If you test ten hooks and keep three tries per hook at 15 seconds each, the footage bill is about $45.60 before you have edited, captioned or exported anything. That is cheap next to a human creator, and it is also a number you can budget, which is the point of reading the real rate.

The page lists 720p. We did not find a separate 1080p or 4K multiplier on the pricing page, so check the resolution you plan to ship before assuming the same rate.

Is Veo 3 still available anywhere

Not through the Gemini API model names that older guides use. Developers who built on veo-3.0-generate-001 or the 3.1 preview names have been told to migrate. Consumer and enterprise Google products can differ from the developer API, and plans there are priced separately, so a Gemini app subscription is not a substitute for the API rate above when you need to estimate cost per ad programmatically.

If your workflow depended on a Veo-specific look, plan a test week rather than a swap. Omni Flash is described as handling multi-input reasoning, character consistency and multi-turn editing, which is a different feature set from a straight text-to-clip model, so outputs will not match your old clips shot for shot.

How to budget AI video for ads after Veo

1. Decide the finished length. Paid social ads that run 15 to 30 seconds put you at $1.52 to $3.04 a generation.
2. Multiply by attempts. Use three for a tight brief and five for anything with a person speaking to camera.
3. Add the work around the footage: script, captions, 9:16 crop and export. Raw model output is not an ad until those are done.
4. Re-read the pricing page before each campaign. Google moved a flagship video model out of the API in a few months, and the Gemini pricing page itself shows rates that change on 1 December 2026 for other models.

Before you decide what to generate, it helps to see what competitors are already running. A tool that pulls competitor ads from Meta, Google and TikTok shows you the hooks to beat, so your first generations are not blind guesses.

Where AdsGen fits

A video model answers "can I generate this shot". It does not decide what the ad should say or who it should say it to. AdsGen takes a product URL and returns up to 20 UGC scripts across 6 angles per batch, then renders selected scripts as 9:16 videos with burned-in captions. Plans start at $49 a month billed yearly.

AdsGen plans include few video renders. Starter includes 1 a month, Growth 2 and Scale 6, so a raw model API wins on clip volume. AdsGen makes sense when the bottleneck is knowing which hook to spend renders on. The AdsGen pricing page lists the allowances, and our Kling AI pricing breakdown prices the main alternative per finished ad.

Veo 3 pricing FAQ

How much does Veo 3 cost?

Veo 3 is no longer sold through the Gemini API. Veo 3.0 shut down on 30 June 2026 and Veo 3.1 Preview on 22 October 2026. Google's current video model, Gemini Omni Flash, costs about $0.10 a second of 720p video, so a 15-second clip is roughly $1.52 per generation.

Is Veo 3 free?

There is no free tier for Gemini Omni Flash, the model that replaced it. The pricing page marks the free tier as not available and the model as paid only. Budget real money for testing, and use the attempts table above to size it.

What replaced Veo 3 in the Gemini API?

Gemini Omni Flash, model name gemini-omni-1.1-flash. Google's deprecations page lists it as the recommended replacement for veo-3.0-generate-001, and the video guide tells Veo 3.1 Preview users to migrate to it or to the GA models on the Gemini Enterprise Agent Platform.

How much does a Gemini Omni Flash video cost per second?

Output is billed at $17.50 per million video tokens, and Google says one second of 720p video uses 5,792 tokens. That works out to about $0.10 a second. Input tokens at $1.50 per million are billed separately and are small for a text prompt.

What does a 30-second AI video ad cost to generate?

About $3.04 for a single 30-second generation at 720p on Omni Flash, or roughly $9.13 for three attempts. That covers footage only. Scripting, captions and 9:16 export are extra, which is the part AdsGen handles from a product URL.