10-SKU AI video ad pilot before Black Friday: priced line by line
A 10-product video ad test costs $22.74 on Sume: image edit, 5 s Seedance 2 clip, $0.20 captions and a $0.10 Timeline end card, plus $3.78 for two retakes.

A ten-SKU pilot of AI video ads costs about $22.74 on Sume at 720p, or about $26.52 if two clips need a retake. Black Friday is Nov 27, 2026, so a pilot run in the week of Oct 5 leaves time to learn which product photos survive motion before you commit a full catalog. The numbers below come from the Sume pricing code and the Sume pricing page, read 2026-10-05; billable price is the provider price times 1.25.
The four lines per SKU
Each SKU gets the same short chain: an image edit that cleans the product photo, one image-to-video clip, captions burned in from authored cues, and a Timeline render that appends an end card. Image-to-video on Sume uses POST /v1/videos with a first_frame entry in frame_images, described in the videos docs.
| Step | Model or endpoint | Unit price | Per SKU |
|---|---|---|---|
| Clean the photo | gpt-image-2.5 edit, 1 reference, high quality | $0.0835 | $0.0835 |
| Motion clip | Seedance 2, 720p, 9:16, 5 s | $1.8900 | $1.8900 |
| Captions | POST /v1/video-captions, cues | $0.20 per job | $0.2000 |
| End card render | Timeline 1.0, under 1 minute | $0.10 per minute | $0.1000 |
| Total | $2.2735 |
Ten SKUs and two retakes
Ten SKUs at $2.2735 is $22.735. If two clips come back with a bent label or a warped hand and you regenerate only the motion step, add 2 x $1.89 = $3.78 for $26.515 total. Captions and Timeline are deterministic, so they are not repeated unless the copy changes.
The caption line works on silent clips only if you send cues with text, start and end; without them the job fails with caption_no_speech, as the caption docs state. The caption price covers videos of 60 seconds or less.
What the pilot should decide
Use the pilot to choose a motion tier, not to prove a campaign. Re-price the same ten SKUs on Seedance 2-mini at 720p ($0.945 for 5 s) or Wan 3.0 at 720p ($0.625) and compare the clips side by side.
- Keep photos that survive a 5 s clip without label drift; those SKUs move to the full run.
- Move SKUs that fail twice to a still-image ad instead of paying for a third clip.
- Write the per-SKU cost into your sheet so the full-catalog multiplication is arithmetic, not a guess.
For the catalog-scale version of this arithmetic see the four-phase Black Friday budget, and for how rejects change the total see the reject-rate budget.
Sources
Related posts
More in Pricing
- Text-to-speech cost per spoken minute: Sume, ElevenLabs v4 and MAI
Convert per-character TTS prices to dollars per finished minute, assuming 900 characters of script. Sume is $0.0428 a minute; vendor list prices compared.
- Transcribe 500 thirty-second ad clips: MAI Streaming hours vs Sume
Five hundred 30-second ad clips are 15,000 seconds of audio: $2.25 on MAI-Transcribe-2-Streaming, $2.50 on Sume STT. Short files suit one job each.
- Translated script length changes the TTS bill: count characters
Sume TTS bills each character at $0.0475 per 1,000, so a translation that runs 25 percent longer costs 25 percent more. Count characters before you speak.
- One TTS plus lip-sync clip: Sonic voice and H3 Max at 768p
A 5 second talking clip on Sume costs $0.51 before the platform fee: $0.01 for the Sonic voice and $0.50 for H3 Max at 768p. A 14.8 second clip is $1.51.
Written by Sume