Veo 3.1 API

Six tiers carrying two prices: $0.0067 per second on fast and lite, $0.1667 on the full model. Billing counts the seconds of the clip that came back, and resolution does not enter the sum at all.

Six tiers, two prices

Veo 3.1 occupies six rows in our catalog and the price list folds them into two groups. Fast and lite cost the same, and inside either group hd and fullhd cost the same as well. Choosing a resolution is free, while moving from fast to the full model multiplies the bill by 25.

Which makes the opening question not "which tier" but "do you need the full model at all". Eight seconds on fast run $0.05, the same eight seconds on the full model run $1.33, and when clips come in batches rather than one at a time, that gap settles your budget long before anyone gets to arguing about picture quality.

What separates fast from lite is not something our data can answer. One price, one set of request fields, different ids. As with the nano-banana pair, only your own test decides it.

Tiers and prices

You pay for what came back, per second of finished video. Order eight seconds, receive six, pay for six. While the job runs, a hold sits on the balance, sized by the duration you asked for or by the maximum eight seconds when you left duration out. The hold comes off if the clip never arrives.

Veo 3.1 at other aggregators

Prices read on 22.08.2026 from each service's own public pages. For most of them the top of the range is the with-audio rate, while our catalog carries no separate audio tier.

Our top rate lands below the base rate at fal.ai, replicate, wavespeed and segmind. Our bottom rate belongs to a different range altogether, though the only honest comparison for it is against their fast tiers.

How to call it

POST /v1/videos, same key as everything else. With "wait": true the gateway holds the connection and hands back the finished clip. Without it, you get a job id straight away.

curl https://api.teamtoken.store/v1/videos \
  -H "Authorization: Bearer $TEAMTOKEN_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "veo-3.1-fast-fullhd",
    "prompt": "camera glides along an empty night platform, snow in the lamplight",
    "duration": 8,
    "aspect_ratio": "16:9",
    "wait": true
  }'

What the request takes

Duration is discrete here. Asking for seven seconds returns an error listing the values that are allowed. Resolution comes from the model you picked, no field will change it, and it makes no difference to the price anyway.

Waiting, and what arrives

With "wait": true the gateway polls the provider for up to a minute and a half. If it finished in time, you get completed with the duration, the charged amount and a link to the file. If it did not, you get a 202 and a job id, and the result comes later from GET /v1/videos/{id}.

The link in the response points at our domain rather than the provider's CDN, and that is deliberate. The provider changes its delivery host more often than anyone would like, and your link should not depend on that.

A failed generation costs nothing. The response carries a zero cost_usd and the hold comes off.

Veo 3.1 and Veo 2

Veo 2 is in the catalog too, at $0.0333 per second. Dearer than the fast tiers of 3.1 and 5 times cheaper than the full model. It is the previous generation and we are not hiding that: when the difference does not show up in your work, the full model buys you nothing.

No gallery here, but there is something to test

No clips of ours here. What this page claims concerns price and resolution, and three calls settle it, two of which cost pennies.

The full set comes to $1.44, taken off your balance rather than run in some demo mode. If only the main question interests you, the first 2 calls at $0.11 answer it.

Questions we get