AI Video API Pricing: Veo 3.1 vs Runway Per-Second Rates
- Google Veo 3.1 audio rates differ by tier and resolution; Runway converts model credits at $0.01 each.
- Input charges, output formats, permitted duration and minimum charges can change a comparison.
- Separate successfully billed output from clips accepted by your creative review; no measured acceptance rate is assumed.
AI video generator API pricing depends on the provider, model, resolution, audio, input and requested output format. A cheap per-second row is useful only when it describes the request you can actually submit. Below are selected current Google Gemini API and Runway Dev rates read on October 11, 2026, followed by a method for budgeting accepted footage. This is a bounded comparison, not an exhaustive market ranking or a live checkout quote.
Price the valid request first, then count the clips that actually pass your review.
Google Gemini API video pricing: tier and resolution
Google’s current Veo 3.1 pricing lists video with audio per second in USD. Standard costs $0.40 at 720p or 1080p and $0.60 at 4K. Fast costs $0.10 at 720p, $0.12 at 1080p and $0.30 at 4K. Lite costs $0.05 at 720p and $0.08 at 1080p, with no 4K output support. These rows are not interchangeable settings of a single flat-price request.
An eight-second 720p Fast generation therefore has $0.80 of listed output cost; an eight-second Standard generation at 720p has $3.20. At 1080p, eight seconds of Fast has $0.96 of listed output cost. These are rate multiplications for the specified outputs, not invoices from our own paid experiment or evidence that the models produce equivalent creative results.
Google also lists Gemini Omni Flash using token-based pricing, with an effective video-output rate around $0.10 per second at 720p, plus separately listed input and text-output rates. Do not turn that effective output rate into an all-inclusive flat-price API call. The useful decision is which supported request produces the needed asset, followed by the provider’s applicable billing information.
- Google Veo Fast: 720p $0.10/s; 1080p $0.12/s. Veo 3.1 with audio; verify valid duration and resolution.
- Google Veo Standard: 720p/1080p $0.40/s. Veo 3.1 with audio; 4K has a separate $0.60/s rate.
- Runway Gen-4 Turbo: $0.05/s base model rate. Image-to-video; a source image and valid request are required.
- Accepted-clip cost: Actual billed total / accepted clips. Creative rejection is not the same as generation failure.
Runway Dev pricing: credits and model-specific extras
Runway’s current pricing sets a developer credit at $0.01 and lists Gen-4 Turbo at five credits per generated second, or $0.05 per second. Its model index identifies this as image-to-video. Gen-4.5 is listed at twelve credits per second, or $0.12, and supports text-to-video and image-to-video. The lower base rate does not establish that a source-image workflow and a text-only workflow are interchangeable.
The same pricing page adds charges for supported professional formats on Gen-4.5 and Aleph 2.0: ProRes or PNG sequences add five credits per second. Listed HDR and ten-bit profiles carry larger surcharges, with an additional distinction for outputs above four megapixels. Compare the actual format your editor requires rather than describing the base generation price as the complete cost of every deliverable.
Other rows use different meters. Seedance 2.5 charges separately for output video and input or reference video, with resolution tiers and a minimum per generation. Model Router requests are billed at the selected model’s standard rate, and response metadata reports the realized cost. A router is not evidence of one permanent rate. Taxes may apply, and these developer credit prices should not be conflated with a consumer subscription allowance.
Validate duration, input and delivery before comparing
Google’s Veo model guide lists four-, six- and eight-second generation lengths with conditions. The current 1080p and 4K settings require eight seconds, and extensions and reference-image workflows have additional restrictions. Confirm the model and input combination before making an example. A hypothetical three-second request should not be priced as if the documented endpoint necessarily accepts it.
For a Runway model, inspect its current endpoint parameters and your source-image or text workflow before ordering. This article does not assume that every API model accepts every duration, ratio or audio setting. Include the cost of preparing a required input image when evaluating the work, even if that preparation happens in another tool rather than on the video provider’s invoice.
Save the request settings, task identifier, billing response and downloaded output together. Google’s Veo guide says generated videos are retained on its server for two days and must be downloaded to keep a local copy. That is a delivery condition, not extra footage in a pricing plan. A returned task or temporary output URL should not be counted as an accepted file in the customer’s final handoff.
Estimate the cost of usable footage without inventing a success rate
Separate technical generation status from creative acceptance. Google’s pricing notes that some audio-processing issues can prevent video generation and that it charges when a video is successfully generated. A generated clip with the wrong product detail is a different issue: the fact that your editor rejects it does not establish a refund or a waived charge. Apply each provider’s actual billing conditions to failures.
Consider a hypothetical eight-second 720p Veo Fast brief. If ten successfully generated clips incur the listed $0.80 output cost each, the generation output total is $8. If creative review accepts two, that hypothetical output expense is $4 per accepted clip, before other work. Those counts are assumptions illustrating the denominator; they are not our observed acceptance rate or a forecast for your project.
Use actual totals when the experiment runs: include applicable input and format charges, then divide by accepted clips or accepted seconds under a stated review rule. Keep labor, assets and assembly costs visible alongside API charges. Comparing a $0.05 base row with a $0.40 audio row alone cannot establish the cheapest finished ad, especially when the requested inputs and deliverables differ.
When a finished-video workflow is a better comparison
An API is suitable for a developer integrating a specific generation step. A creator who needs an edited video with spoken copy, subtitles and a reviewed sequence should also compare the work after that step: scripting, selecting footage, editing and final export. API integration time and product workflow costs answer different questions. Neither is automatically a better purchase without the required output.
NoobClaw’s documented example is a local script-to-video export, not a Google or Runway benchmark. A 441-character script produced a 32.6-second, 1080-by-1920, 30fps file with observed audio and subtitles. Generic footage diverged from the subject, and voice quality and asset rights were not approved for publication. It establishes an inspectable production result and remaining review work, not a comparative API cost, acceptable synthetic ad or platform delivery.
If your immediate task is a finished creative draft, inspect that workflow and current product offer. If you require a provider API, select a valid model configuration and run a small authorized paid test with the actual billing record and acceptance rule. This article has not executed that purchase. Preserve which facts come from the catalogue, which figures are arithmetic and which results you truly observed.
For the related content and workflow decisions, read AI UGC Creator: Jobs, Deliverables and Disclosure and Postiz Open Source vs Paid: Plans, AI and Hosting Costs. Each review has its own evidence and scope.
NoobClaw publishes this article. Its script-to-video guide and public creative execution record document a local video result, not external-platform publication. Check the current product offer against the supported action you need.
Frequently asked questions
Which AI video API is cheapest?
This article compares selected Google and Runway rates, not the whole market. Select a valid input, audio, resolution, duration and format before comparing listed base rates; accepted-output costs require an actual review sample.
Is a per-second rate the total price of an API request?
Sometimes it is only one part. The cited catalogues also include input/reference-video charges, format surcharges, minimums and token-based video output. Check the model-specific billing rule.
Did NoobClaw measure these APIs’ usable-clip cost?
No. The prices are current catalogue facts and the accepted-clip example is hypothetical. The separate public NoobClaw record documents one local export with remaining review limitations.
Sources
Official material read on October 11, 2026. Prices identify their billing term; vendor allowances are not measured publishing results. Hypothetical inventory and cost examples are labeled; the public local-production record has its own limitations.