Seedance 2.5 draft mode: a 480p-then-1080p loop on Sume
Higgsfield added Seedance 2.5 Draft Mode. On Sume there is no draft model: render at resolution 480p, pick a take, then re-request at 1080p.

Sume has no separate draft mode for Seedance 2.5. You get the same effect by sending resolution: "480p" for the preview, choosing the take you like, and sending a new request at 1080p. Each request is a normal paid job.
The Higgsfield changelog (Sep 22, 2026) describes its Draft Mode as previews at 480p, then a final at 1080p on Seedance 2.5, and says it is unavailable in the US due to provider constraints. Sume's side is from the Video Router docs and catalog code, read 2026-09-30.
What does Sume let me change between draft and final?
| Item | Higgsfield Draft Mode | Sume `seedance-2.5` |
|---|---|---|
| Preview resolution | 480p | 480p |
| Final resolution | 1080p | 1080p (also 720p) |
| Length | Not stated in the entry | 4 to 30 s |
| Draft as a separate mode | Yes | No; only resolution changes |
Will the final match the draft?
Do not assume it will. The Sume docs describe no seed field and no promise that a second request reproduces the first. Treat the 480p render as a check of composition, motion and pacing, and expect the 1080p render to be a new generation of the same prompt.
How do I run the loop safely?
Use a distinct Idempotency-Key per request, as the docs' example does with video-router-001, so a retry of the preview does not create a second paid job. A reused key means the same request, so give the 1080p final its own key. For paid tools through MCP, the Sume skill guidance says to preflight the exact request with the tool's dry run, which previews admission and cost without submitting the job.
curl -X POST https://api.sume.com/v1/video-router/generate \
-H "Authorization: Bearer $SUME_API_KEY" \
-H "Content-Type: application/json" \
-H "Idempotency-Key: draft-480p-001" \
-d '{
"model": "seedance-2.5",
"prompt": "A vertical UGC-style product clip on a desk, natural light",
"resolution": "480p",
"duration": 8,
"aspect_ratio": "9:16",
"mode": "async"
}'
# final: same body with "resolution": "1080p" and Idempotency-Key: final-1080p-001How much does the draft save?
Billing is by output, and rates differ by resolution, so read the current numbers from the catalog row or the Seedance 2.5 price page instead of assuming a ratio. The check that matters is whether previews let you skip full renders you would have discarded.
Sources
Related posts
More in Use cases
- Seedream 5.0 pro bbox and point tags vs region edits on Sume
Seedream 5.0 pro edits a region from <bbox> and <point> tags on a 0-999 grid. Sume lists other Seedream ids; use mask_url on ChatGPT Image 2.5 for regions.
- Shopify's recommended 2048 x 2048 product image with gpt-image-2.5
Shopify.dev recommends 2048 x 2048 px for product images. That is a valid custom image_size on Sume's gpt-image-2.5: 2048 is a multiple of 16 and 4,194,304 px.
- Black Friday homepage hero image for Shopify: one set, many sizes
Shopify's BFCM checklist says to put seasonal motifs in the hero image and extend it across channels. Render one reference to each channel's ratio with Sume.
- Black Friday teaser video for a Shopify store: a batch plan
Shopify's BFCM checklist says to send warm-up emails with hints and teasers. Make one teaser per product from a sheet with a Sume Format bulk run.
Written by Sume