100 product photos to 5-second AI clips: cost by model on Sume
100 product photos as 5-second image-to-video clips cost $38.00 to $289.00 on Sume depending on model. Ten rows, per clip and per batch.
A 100-photo catalog turned into 5-second image-to-video clips costs $38.00 on H3 at 768p and $289.00 on Seedance 2.5 at 720p on Sume, with the other rows in between. That is one clip per photo, one take each, no retries.
Each photo goes in as the first_frame of frame_images. The prices below are the per-clip billable total (list times 1.25, rounded up to the cent per job) times 100, so the rounding is applied before the multiplication.
Cost of the batch
Seedance rows are per video token at 16:9 and 720p; the other rows are per second. Sorted from cheapest to most expensive.
| Model and tier | Per clip | 100 clips |
|---|---|---|
| H3 768p | $0.38 | $38.00 |
| H3 Max 768p | $0.50 | $50.00 |
| Wan 3.0 720p | $0.63 | $63.00 |
| Gemini Omni Flash 1.1 720p | $0.63 | $63.00 |
| Kling 3 audio off | $0.70 | $70.00 |
| Seedance 2 mini 720p | $0.95 | $95.00 |
| Kling 3 audio on | $1.05 | $105.00 |
| Seedance 2 fast 720p | $1.52 | $152.00 |
| Seedance 2 720p | $1.89 | $189.00 |
| Seedance 2.5 720p | $2.89 | $289.00 |
What changes the number
A retry is a second job and a second reserve, so the realistic budget is the per-clip price times your takes per keeper. If two takes in three are usable, add half again. The cost-per-keeper post shows the arithmetic, and failed jobs explains what a provider failure costs.
Audio matters on one row only. Kling 3 is $0.14 a second silent and $0.21 with audio, so audio on adds 50 percent. Omni, Wan, H3 and H3 Max include audio in the rate.
Two cheaper habits
- Preview at low resolution first. Wan 3.0 at 480p is $0.32 a clip, so 100 previews are $32.00; promote only the keepers.
- Use an
Idempotency-Keyper photo (for example the SKU) so a network retry does not submit a duplicate job. - Submit in batches and poll, or pass a
callback_urlover HTTPS and let results come to you.
Which row for product photos
Photos have a fixed aspect, so check it first. Kling 3 takes the aspect from the start image, and Omni accepts 16:9 and 9:16 only. Wan 3.0 offers more ratios. For a plain catalog loop, Wan 3.0 or H3 Max at the lower tiers keep the batch cheap, and Seedance 2.5 is the premium choice when a single hero shot matters more than the batch average.
Sources
Related posts
More in Use cases
- 100 SKUs in 5 languages: 500 caption jobs, $100.00 and a key scheme
Captioning one clip per SKU in five languages is 500 Sume caption jobs at $0.20, $100.00; key each job sku-language so a restarted batch never pays twice.
- 12 AI Reels a month, 6 seconds, 3 takes each: cost by model on Sume
12 six-second vertical Reels with three takes each is 36 jobs. It costs $16.20 to $124.92 on Sume depending on model.
- 12-second AI video with first and last frame and sound under $2
Four Sume rows hold a 12-second clip with first and last frame and audio for under $2: Wan 3.0 and H3 Max at 480p or 720p/768p, and H3 at 480p or 768p.
- 160x600 skyscraper ad image: 3.75:1 is past GPT Image 2.5's 3:1 cap
A 160x600 ad is 1:3.75, taller than the 3:1 limit on GPT Image 2.5 custom sizes. Render 576x1728 (1:3) and crop 20% of the width in Pillow. Math included.
Written by Sume