Quote the generated footage behind a finished AI video
Update the endpoint first, then calculate generated footage, audio settings and rejected takes before quoting a finished video.

A sixty-second finished video can consume several minutes of generated footage. The unit on a provider's price page is therefore the beginning of a production estimate. You also need to know which endpoint accepts the request and how much of each returned clip you expect to keep.
Prices below were checked on November 14, 2026. Older Veo 3.1 preview price comparisons describe an earlier service configuration. Google's current documentation marks those Gemini API preview endpoints as shut down on October 22, 2026. A quote built around an unavailable endpoint needs a replacement model before its arithmetic becomes useful.
Verify the service before the rate
Google's current video generation guide directs Gemini API users to Gemini Omni Flash and points to generally available Veo models on its separate enterprise platform. That distinction belongs in a purchasing decision. A model family name can appear through several services whose features and terms differ.
Record the provider and the exact endpoint in your estimate. Include the output resolution and whether you need generated audio. A price copied under a bare model name loses the details another person needs to reproduce the quote. It can also outlive the endpoint that originally made it accurate.
For an existing integration, check the deprecation notice before planning a repeat order. A stored request may require a different schema after migration. Allocate time to verify the replacement output against the brief. Endpoint availability and acceptable visual behaviour are both conditions of delivery.
Translate the documented meter carefully
Google's Gemini API pricing page currently describes Gemini Omni Flash video output through token billing. At the documented standard rate, 720p output works out to approximately seven cents per second. That is an effective output rate derived from the specified token consumption, with input charges accounted for separately.
For that output component, sixty generated seconds are approximately $4.17 using the documented token calculation. Rounding the per-second estimate to seven cents produces $4.20. Keep the precision consistent when communicating an estimate. Neither figure is a promise that a complete one-minute advertisement will cost that amount.
The same table assigns different effective costs to higher output tiers. Select the tier required for delivery before comparing providers. A low-resolution draft can help settle a composition, although a later higher-resolution result still needs review. Budget the final pass explicitly when the workflow requires it.
Keep the reseller attached to its quote
The fal endpoint for Kling 2.6 Pro text-to-video currently lists $0.07 per generated second with audio off and $0.14 with audio on. It offers five- or ten-second output durations. Those are fal's published terms for that endpoint, checked on the date above.
Six ten-second generations would therefore have listed generation costs of $4.20 without audio or $8.40 with audio. This calculation describes sixty seconds of returned clips. Building a coherent minute from those clips may involve additional generation attempts and editing. Avoid presenting six separate successful requests as a guaranteed single continuous take.
Keep a provider comparison within comparable settings. If one quote includes audio while another requires a separate soundtrack, account for that work in the proposed production workflow. Check the actual endpoint schema before assuming that image input carries the same price as text input. A family resemblance between endpoints leaves that detail unresolved.
Estimate the footage you will discard
Consider a hypothetical brief with six shots, each requiring ten usable seconds. Suppose three generated versions per shot are reviewed before one is selected. That creates 180 generated seconds for sixty accepted seconds. At fal's listed audio-off rate above, the generation component would be $12.60.
The three-attempt assumption is a planning example. Establish your own estimate with a representative trial before relying on it for a fixed-price job. A nearly correct result can still fail because the product changes shape halfway through the clip. Save the rejection reason so that the next brief can address the actual difficulty.
Track useful seconds at the shot level. A ten-second return may contain only four seconds that fit the edit. Counting the entire return as usable footage hides the cost of filling the gap. An editor's acceptance decision gives the estimate a connection to the delivered timeline.
Separate an unsuccessful provider request from a successful return that the editor rejects. The first case follows the service's billing rules. The second can still create a generation charge even though the footage stays outside the edit. Keep both outcomes in your production log. A missing task result should be reconciled with the provider's task status before another attempt is submitted, so the estimate does not lose track of already generated footage.
Build a quote that survives revision
Separate the provider's generation charges from the work of editing the final piece. Describe how many review rounds the quote includes. If the client changes the product reference after approving the first draft, record the extra generation needed for that revision. The source of the cost then remains visible to both sides.
Recheck the live pricing page before sending a final quote. Save the date and the settings alongside the arithmetic. Taxes and any purchased subscription allowance belong in the account-specific calculation. A public list price provides a common reference while your own billing arrangement determines the actual payment.
The useful comparison is the expected cost of delivering this brief through a specified service. Verify endpoint availability before estimating the footage required under the chosen settings. That estimate gives a client enough information to approve the work and gives the producer a basis for reviewing the eventual bill.