AI video cost per deliverable: a rate sheet for finance approval
A cost-per-deliverable sheet built from Sume's published rates: 15-second avatar ad, captioned cut, 6-image set, voiced script, music bed, with a retake factor.
Finance approves deliverables more easily than per-second rates, so price the thing that ships. On Sume's published rates a captioned 15-second avatar ad is about $3.88 for one take, $5.81 with a 1.5 retake factor, and a voiced 1,200-character script is $0.057. The table below builds each line from the rate card read on 2026-10-03.
Everything here is arithmetic on published rates. Plan fees are separate, and the totals exclude tax.
What does each deliverable cost?
Rates: Avatar Video Plus $0.245 per second, caption job $0.20, timeline render $0.10 per output minute as a reserve ceiling, music $0.125 per track, TTS $0.0475 per 1,000 characters, background removal $0.0225 per image.
| Deliverable | Build | One take | With 1.5 retake factor |
|---|---|---|---|
| 15 s avatar ad, captioned | 15 x $0.245 + $0.20 | $3.875 | $5.81 |
| 15 s avatar ad, captioned and timeline-cut | $3.875 + $0.10 | $3.975 | $5.96 |
| 30 s avatar ad, captioned | 30 x $0.245 + $0.20 | $7.55 | $11.33 |
| Voiced 1,200-character script | 1.2 x $0.0475 | $0.057 | $0.09 |
| Music bed | 1 track | $0.125 | $0.19 |
| Six cutout product images | 6 x $0.0225 | $0.135 | $0.20 |
Why a retake factor?
Generation is not deterministic, and a rejected take is a captured job. Failed jobs are refunded; jobs that succeed but are not used are not. A factor of 1.5 means one extra take for every two used. Take the factor from your own last month: used clips divided into captured jobs. Do not borrow a number, because mine is an illustration.
How do you turn this into a monthly budget?
Multiply deliverables per month by the right column and add the plan fee. A team shipping 30 captioned 15-second ads, 30 voiced scripts and 10 music beds is 30 x $5.81 + 30 x $0.09 + 10 x $0.19 = about $179 in usage, about $219 on Pro. The plan table and the Pro workload breakdown are in what a month on Pro costs.
How do you check the spend against the sheet?
Tag each deliverable with the job ids you got at submit and read debited_usd_micros from the scoped usage summary after the jobs settle. Divide by the number of approved deliverables and compare to the sheet. The client cost ledger post shows a small script for it.
If the actual cost per approved deliverable runs above the right-hand column, the retake factor is higher than you assumed. That is the number to take back to the creative team, not the per-second rate.
What is not covered?
Avatar creation at $0.95 each, product-image inputs, the cost of your own review time, and any tax or card fee. The timeline line is a ceiling, so captured cost can be lower. Re-read the rate card before you present the sheet: it is dated and can change.
Sources
Related posts
More in Pricing
- AI virtual try-on cost per image: GPT Image 2.5 on Sume
What one try-on still costs on Sume: the documented token rates, why quality auto reserves the max price, and a script that reads the live endpoint pricing.
- AI voice agent cost per minute: speech-to-text, LLM and TTS stack
A dated cost sheet for the October 2026 voice stack: MAI-Transcribe-2-Streaming, Mercury Voice, MAI-Voice-2.1-Flash, and Sume's file STT and TTS, per minute.
- Change a person in a video with AI: Sume Recast price, not free
Changing a person in a video with Sume's H3 Max Recast is paid per source second: $0.375 at 768p, $0.5625 at 1080p. Costs for 5 to 30 second clips.
- Opus 5.5 costs 20% less than Opus 5: video agent turn math
Opus 5.5 lists $4 input and $20 output per MTok against Opus 5 at $5 and $25. Priced on a 200K-token video-agent turn, and what Sume did with Opus 5.
Written by Sume