Seedance 2.0, Fast and Mini next to 2.5: still callable on Sume
Sume still serves seedance-2, seedance-2-fast and seedance-2-mini by id next to seedance-2.5. 15-second ceiling, per-token prices, 4 s and 5 s job costs.

Yes, Seedance 2.0 is still callable on Sume. The catalog keeps seedance-2, seedance-2-fast and seedance-2-mini next to seedance-2.5. They are explicit picks: sume/auto does not choose them.
The older ids stop at 15 seconds, where 2.5 reaches 30, and they cost less per second of video.
What does each Seedance id cost on Sume?
Seedance is priced per 1,000 video tokens rather than per second. Billed per million tokens, after the 1.25 multiplier, the four ids are $8.75 for Mini, $14.00 for Fast, $17.50 for 2.0 and $26.75 for 2.5 (the 2.5 row uses a different rate at 1080p). Tokens grow with width, height, seconds and frame rate.
| Model id | Longest | 4 s at 720p | 5 s at 720p |
|---|---|---|---|
| seedance-2-mini | 15 s | $0.76 | $0.95 |
| seedance-2-fast | 15 s | $1.21 | $1.52 |
| seedance-2 | 15 s | $1.52 | $1.89 |
| seedance-2.5 | 30 s | $2.32 | $2.89 |
When is an older id the better pick?
If the clip is 15 seconds or shorter and you do not need Seedance 2.5's features, the cheaper ids give the same request shape for less. Mini is the draft row: at 720p a 4-second clip is about a third of the 2.5 price.
What the newer id adds on Sume is the 30-second ceiling (4 to 30 seconds against 4 to 15) and its own token rate. Those are the reasons to pay for it.
- Draft with
seedance-2-mini, final withseedance-2orseedance-2.5. - Same request fields across all four ids, so switching is one string.
- Check
supported_resolutionsper id in the models list before pinning 1080p.
Why does the catalog list them separately?
Each id is a distinct provider model with its own rate and limits, and Sume's contract uses bare ids so a request always says which one ran. Only sume/auto hides the family, and Auto's default is Gemini Omni Flash 1.1.
How do you pin one?
Pass the id in model. The sample drafts on Mini.
import os, requests
r = requests.post(
"https://api.sume.com/v1/videos",
headers={"Authorization": f"Bearer {os.environ['SUME_API_KEY']}"},
json={
"model": "seedance-2-mini",
"prompt": "A paper boat drifts down a rain gutter",
"duration": 5,
"resolution": "720p",
"aspect_ratio": "16:9",
},
)
print(r.status_code, r.json()["id"])Sources
Related posts
More in Models
- Seedance 2.5 on BytePlus excludes the US: check access first
BytePlus lists Seedance 2.5 on ModelArk in supported markets excluding the United States. What that means for a US team, and how to read the Sume catalog.
- Self-host MiniMax H3 or use an API: a decision table
MiniMax released H3 as an open-weight omni-modal model on Jul 31, 2026. A decision table for running it yourself versus calling a hosted video API.
- SF3D makes a GLB in 0.5 s: what it skips
Stability's SF3D is listed at $0.07 on fal.ai, GLB only, about 0.5 seconds. Sume does not carry 3D, so here is what to use for product visuals instead.
- Stable Audio 3 sizes: which one for your job
Stable Audio 3.0 ships Small SFX, Small, Medium and an API-only Large. Match size to job, and see how Sume's Music Router handles length with a prompt.
Written by Sume