Knitwear flat-lay to a moving on-model ad: three colorways for $5.92
One flat-lay, three colorways: three $0.0835 on-model image edits, then three 5 s Seedance 2 clips at $1.89. $5.92 on Sume, plus the check to run between.

A knitwear flat-lay becomes a moving on-model ad in two steps: an image edit that places the garment on a model, then an image-to-video clip that animates that still. For three colorways of one sweater, the cost on Sume is 3 x $0.0835 for the stills plus 3 x $1.89 for 5-second 720p Seedance 2 clips, which is $0.2505 + $5.67 = $5.9205, or $5.92 (prices read 2026-10-05). Knits are a hard case, since stitch pattern and neckline shift when a model moves, so the plan below puts a check between the two steps.
Step one: the on-model still
gpt-image-2.5 accepts up to 16 reference images, per the image docs. Send the flat-lay as a reference and describe the model, the pose and the setting. An edit with one reference at high quality is $0.0835 on Sume; medium is $0.0209 and low is $0.0094, so you can draft cheaply and finish at high. The existing guide to flat-lay to on-model images covers the prompt.
The check between steps
Motion multiplies any error in the still. If the neckline, cuff or pattern is wrong in the image, a clip will carry it for five seconds. Compare the still against the flat-lay at full size before paying $1.89. Reject any still where the garment differs from the product, because a product ad that shows a different sweater is a returns problem.
- Neckline and hem match the flat-lay.
- Stitch pattern and colour are the product's, not a lookalike.
- Hands, cuffs and sleeves look plausible at full size.
- The model's pose leaves room for a slow camera move, not a big turn.
Step two: the clip, and the bill
Pass the approved still as the first_frame on POST /v1/videos with seedance-2, 5 seconds, 720p and 9:16 (videos docs). Ask for one slow move, such as a gentle turn or a dolly in; keep the prompt short.
| Line | Unit price | Count | Total |
|---|---|---|---|
| Image edit, 1 reference, high | $0.0835 | 3 | $0.2505 |
| Seedance 2, 720p, 5 s | $1.89 | 3 | $5.67 |
| Total | $5.9205 | ||
| With one retake per colorway | $1.89 | 3 more | $11.5905 |
If you need the garment fitted to a specific body, rather than a model who wears it, compare virtual try-on and virtual fitting; for a wider shop workflow, see one garment photo to five shop assets. Retakes are the budget risk: one retake per colorway nearly doubles the motion line to $11.34 and the total to $11.59.
Sources
Related posts
More in Use cases
- Landscaping yard transformation clip: Seedance 2.5 first+last frame
Send the muddy yard as first_frame and the finished lawn as last_frame to Seedance 2.5 on Sume: 8 seconds at 16:9 is $2.15 at 480p, $4.62 at 720p.
- Licensed voices only: what an ad team checks before take one
Microsoft says MAI-Voice-2 synthesizes only authorized, licensed voices. Here is the rights checklist to run for any AI voiceover, and what Sume's docs cover.
- LinkedIn Lead Gen Form hidden fields: tag each lead with its variant
A LinkedIn Lead Gen Form allows up to 20 hidden fields. Keep your own index-to-variant table from the Sume bulk queue so every lead traces to one creative.
- LinkedIn Lead Gen Form image 552x200: make it from a 3:1 render
LinkedIn recommends a 552px by 200px form image. Sume has no 552:200 ratio, so ask Ideogram for 3:1, center-crop about 8% of the width, and resize.
Written by Sume