A 30-second social ad
Six approved five-second shots.
Hypothetical 50% acceptance, one minute review per charged attempt and two edit hours. No claim that a given model reaches that acceptance.
Tutorial · Evidence checked 2026-10-15
How to Calculate AI Video Cost per Approved Finished Second
A practical, evidence-led decision guide. Product capabilities and limits are separated from anything that would require hands-on testing.
BenPicks · Original planning calculator
What will your usable AI video actually cost?
A cheap generation is only one line in the budget. Count the takes you pay for, the seconds you keep and the work it takes to approve them.
Hypothetical scenarios, not product tests. Calculator tariffs checked 16 October 2026. Edit the assumptions for your brief; these are estimates, not measured model performance.
Start with a job
Hypothetical 50% acceptance, one minute review per charged attempt and two edit hours. No claim that a given model reaches that acceptance.
Current scenario budget: $113.00$3.7667 / usable second, conditional on approvalBefore tax and excluded costs; input-image creation is excluded unless entered under other costs. Edit the inputs.
Project estimate / entered worksheet
$113.00
USD total, before tax and excluded costs. Input-image creation is excluded unless entered under other costs.
Cost per usable finished second
$3.7667
- Rounded planning attempts
- 12
- Generated seconds charged/modelled
- 60
- Target usable seconds (conditional)
- 30
- Generation / software allocation
- $3.00
- Review work
- $10.00
- Editing work
- $100.00
- Other entered costs
- $0.00
- Planning/worksheet arithmetic only; no video was generated or evaluated by BenPicks.
- Tax, exchange rates, reference-image creation, storage, premium exports, audio post-production and licence fees are excluded unless entered in other costs. Unknown excluded costs are not zero-cost claims.
- Acceptance is your assumption, held equal across tariffs for sensitivity only. Rounded attempts do not guarantee enough usable clips, and per-second cost is conditional on reaching the target.
- Runway Dev API only, standard output at the listed settings; not a Runway web plan, a direct Kling quote or a quality comparison. Prepare a suitable input image separately for Gen-4 Turbo. This is rate-equivalent usage cost, not cash required to buy prepaid credits or a full-project quote.
Six original briefs you can inspect
All workloads, acceptance rates and labour assumptions are invented planning examples. The stopped brief is fictional too.
Eight product loops
Eight accepted loops, each keeping four seconds of a five-second generation.
Hypothetical 25% acceptance for difficult geometry; trim one second from each accepted take. USD 20 additional asset cost is invented.
B-roll for a tutorial
Twelve approved inserts, keeping three seconds from each generated clip.
Hypothetical 60% acceptance and three editing hours. The 36 useful generated seconds are not the duration of the entire tutorial.
A campaign in two formats
Six accepted deliverable clips per format, twelve across both formats; the underlying ideas may overlap.
Assumes separate generations for two supported aspect ratios, not free reframing. 40% acceptance; five edit hours and USD 30 extra cost are assumptions.
Two separate language versions
Six approved shots per version, no reused generated footage assumed.
Hypothetical 50% acceptance and four edit hours. USD 40 covers a partial assumed external language/audio workflow, not a vendor dubbing quote; actual needs may cost more. No language accuracy or lip-sync claim.
A stopped brief: nothing approved
Twenty charged five-second attempts, zero approved usable clips.
Entirely fictional record to demonstrate sunk spend. No cost per approved second can be calculated. Review/edit effort remains in the total.
Selected API tariffs, with their billing surface
Six tariff rows on one platform: Runway Dev. This is not a list of six providers or a ranking of voice, image or video quality. Models and audio settings differ, so equal acceptance is a sensitivity assumption, not equivalent output.
| Model / setting | USD / generated second | Settings | Permitted seconds |
|---|---|---|---|
| Gen-4 Turbo | 0.05 | Image-to-video; standard 720p; no audio priced; suitable input image needed | 2, 3, 4, 5, 6, 7, 8, 9, 10 |
| Kling 2.5 Turbo Pro | 0.12 | Text/image-to-video; standard 1080p; no audio priced | 5, 10 |
| Kling 3.0 Standard — no audio | 0.09 | Text/image-to-video; standard 720p; audio=false | 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15 |
| Kling 3.0 Standard — audio | 0.13 | Text/image-to-video; standard 720p; audio=true | 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15 |
| Kling 3.0 Pro — no audio | 0.12 | Text/image-to-video; standard 1080p; audio=false | 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15 |
| Kling 3.0 Pro — audio | 0.17 | Text/image-to-video; standard 1080p; audio=true | 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15 |
Runway API pricing · Input settings · Duration changelog · Failed-task billing distinctions
Credits convert at USD 0.01 each on this API. Standard output only: no premium format, upscale or additional asset cost included. Web-plan balances and direct-provider prices are separate. In record mode use one billed duration throughout; mixed lengths, refunds and invoice reconciliation require your actual provider records.
The calculation, not a claim about model performance
Planning attempts = required clips ÷ assumed acceptance fraction, rounded up. This is a budget convention, not a confidence level or a guarantee. In record mode, use your entered charged count instead.
Generated seconds = attempts × billed seconds per attempt. Usable seconds = required or approved clips × useful seconds kept. The unit cost divides software, review, editing and other entered costs by usable seconds. Zero usable seconds means an undefined unit cost.
Subscription mode counts the project allocation and separately charged extras once; API usage is not added to those same included credits. The two methods answer different questions.
Use this for generated footage. Avatar lessons and repurposing existing footage have different units. The surrounding guide explains the full method. Here, “usable” seconds mean the approved finished seconds kept after review.
Download the assumptions, calculations and reproduction files
Download the complete calculation kit (ZIP)
- scenario-inputs.csv
- scenario-results.csv
- acceptance-sensitivity.csv
- public-rates.csv
- record-template.csv
- rates.json
- scenarios.json
- sources.json
- calculator.cjs
- reproduce.cjs
- METHOD.md
Copy all reproduction files into one folder. The local reproduction utility uses only the downloaded calculator and input tables; no API key, account or provider call.
Original planning analysis · Version 18 October 2026
A generation rate is not the cost of an approved second.
The same fictional 30-second ad reaches a USD 126 budget at 25% assumed acceptance and USD 106.50 at 100%. Fixed editing work remains in both cases. Acceptance is an input, not a measured success rate for any model.
| Assumed acceptance (%) | Charged attempts | Generated seconds | Usable seconds (by assumption) | Generation USD | Review USD | Editing USD | Total USD | USD per usable second |
|---|---|---|---|---|---|---|---|---|
| 25 | 24 | 120 | 30 | 6.00 | 20.00 | 100.00 | 126.00 | 4.2000 |
| 50 | 12 | 60 | 30 | 3.00 | 10.00 | 100.00 | 113.00 | 3.7667 |
| 75 | 8 | 40 | 30 | 2.00 | 6.67 | 100.00 | 108.67 | 3.6222 |
| 100 | 6 | 30 | 30 | 1.50 | 5.00 | 100.00 | 106.50 | 3.5500 |
Assumptions and exclusions. Retained Runway Dev API Gen-4 Turbo tariff: 5 credits/second at USD 0.01/credit, checked 16 October 2026. Standard 720p image-to-video, suitable input image required, no audio priced. Six five-second approved clips; one review minute per charged attempt at USD 50/hour; two editing hours at USD 50/hour. No free credits are assumed in this scenario. Tax, FX, reference-image creation, storage, premium exports, audio, rights and other excluded costs are not included. This is allocated usage, not the minimum prepaid purchase. Rounded attempts do not guarantee delivery. On the page and figure, USD money amounts are displayed to two decimal places, and USD per usable second to four; CSV and JSON retain full precision for reproduction. Totals use the unrounded values.
Runway developer API pricing · Original reproduction kit
Download the figure (SVG) · Download the data (CSV) · Download the method and assumptions (JSON)
Cite this analysis: BenPicks. A generation rate is not the cost of an approved second. Version 2026-10-18. https://benpicks.com/guides/how-to-calculate-ai-video-generation-cost/#cite-video
Use the existing calculator above to change the assumptions. These figures are not forecasts of customer results.
AI video pricing becomes misleading as soon as “one generation” is treated as one deliverable. A five-second result can consume different credits by model, resolution, audio mode or platform. It may also be rejected, regenerated, extended, edited and reformatted before one second reaches a customer.
The useful unit is cost per approved finished second. It connects the vendor meter to the asset the business can actually publish.
Quick answer: price the approved result
Use this guide for generated footage. For presenter lessons use the avatar-cost method; for cutting existing footage use the repurposing-cost method. They measure different jobs and should not share a single cost ranking.
| Example input | Calculation | Result |
|---|---|---|
| Generated / approved seconds | 600 / 120 | 5 generated seconds per approved second |
| Hypothetical generation rate | USD 0.10 × 5 | USD 0.50 per approved second before labour |
| Same rate, only 60 seconds accepted | USD 60 / 60 | USD 1.00 before labour |
Runway's developer documentation currently prices API credits at USD 0.01 each, with consumption varying by model and mode; tax may apply. That is a credit purchase rate, not a universal per-second price or the price of its web subscription. Record the model row and billing surface together (checked 15 October 2026).
Count each expense once: use either the subscription allocation or the value allocated to consumed included credits, then add separately charged overages. Do not add both for the same allowance. With zero approved output, report total spend and an undefined unit cost. Use LTX Studio versus Filmora when deciding between shot generation and an editing-first purchase.
Build the denominator from the campaign brief
Record the required finished seconds by format, not just the number of videos. A launch might need:
- one 20-second hero film in 16:9;
- three six-second cut-downs in 9:16;
- six four-second product loops in 1:1;
- two nine-second message variants.
That is 80 required finished seconds before alternate takes. Keep a separate count for each format if cropping or regeneration is required. Do not assume one output can be reframed without visual loss.
Translate the provider meter without pretending credits are comparable
For each exact platform/model/mode, capture:
| Input | Example of the evidence required |
|---|---|
| Price paid | invoice or current checkout, including billing period |
| Included allowance | credits, compute units or seconds |
| Model consumption | credits per second or per generation at selected settings |
| Duration/resolution | the exact deliverable the rate covers |
| Expiry/rollover | official plan rule |
| Failed-job treatment | whether credits are restored and under what condition |
| Web/API separation | separate balance, rate card or contract |
Credits are internal meters. Equal credit balances from different providers have no shared economic meaning. Convert each to currency only after the plan price and eligible allowance are known.
Calculate submitted generation cost
When a provider charges per second:
submitted generation cost = total seconds requested × price per generated second
When it charges credits:
credit value = eligible plan price / usable plan credits
submitted generation cost = credits consumed × credit value
The word “eligible” matters. If a plan bundles stock assets the team already needs, finance may allocate only part of the subscription to video. If a Runway allowance is shared across a workspace, use the actual credits consumed by the project, not the whole plan or a fictional equal split.
For annual pools, match the period. Dividing an annual allowance by 12 is useful for steady-state planning but can conceal a seasonal campaign's ability to consume a large share at once.
Add rejection and revision
Define:
technical pass rate = outputs without corruption or platform failure / outputs completed
creative acceptance rate = outputs approved / outputs reviewed
submitted seconds per approved second = submitted generated seconds / approved finished seconds
If 600 generated seconds produce 120 approved seconds, every approved second carries five generated seconds of cost. At an illustrative rate of $0.10 per generated second, that contributes $0.50 per approved second before editing. This is a hypothetical rate, not a current provider quote.
Separate platform failures from creative rejection. A malformed file and a legally unusable likeness are different problems even when neither ships.
Add the labour that cheap generation creates
Track prompt preparation, queue monitoring, review, rights checks, editing, sound, typography, export and project management.
labour cost = hours by role × loaded hourly rate
total approved-second cost = (generation + labour + storage + required seats + rights/licensing) / approved finished seconds
Illustrative budget, not a vendor quote:
- plan allocation: $120;
- overage: $80;
- 8 creative hours at $55: $440;
- 2 legal/brand review hours at $90: $180;
- 160 approved seconds.
Total cost is $820, or $5.13 per approved second. Reporting only the $200 generation spend would hide 76% of the cost.
Compare a creative UI and an API separately
An API adds engineering, storage, retries, moderation handling and observability. A creative UI adds seat, review and handoff costs. Do not transfer the web plan's credit value into the API row unless the official documentation says the balance is shared.
For automated systems, add:
API operating cost = generation + orchestration compute + storage/egress + human exception review
Measure retry amplification: submitted jobs divided by unique approved assets. A silent retry loop can dominate spend.
Run a 20-output calibration before forecasting a year
Use one representative brief and 20 attempts per selected model. Preserve prompt, input rights, settings, credits, duration, result, rejection reason and review minutes. Calculate the observed acceptance rate and use a conservative confidence range rather than treating 20 outputs as certainty.
Repeat the calibration for the hardest content class: people, product geometry, text, continuity or camera motion. An average across easy abstract clips will understate campaign cost.
The worksheet output procurement needs
The final sheet should show:
- currency per submitted generated second;
- submitted seconds per approved second;
- labour per approved second;
- total cost per approved second;
- monthly capacity at the selected plan;
- the first limit reached: credits, seats, queue, rights or reviewer time;
- unknowns that require checkout, contract or hands-on evidence.
That turns a model demo into an auditable budget.
Primary next step: Compare AI video tools by their real billing boundary and calculate one campaign before choosing a plan.