BytePlus Seedance 2.5 plans: how many 30-second 480p clips fit

BytePlus's three Seedance 2.5 plans cover roughly 4 to 16 thirty-second 480p clips on Light, 8 to 33 on Production and 83 to 333 on Premium, and each is valid for 3 months. Those clip counts come from dividing the vendor's own "approx." seconds of 480p video by 30; the page shows a range for each plan because usage is priced per token and varies with resolution and input mode. On Sume there is no plan to buy for this: you submit seedance-2.5 jobs from your workspace balance.

What do the three plans include?

The BytePlus Seedance 2.5 page, read 2026-10-03, lists Light, Production and Premium plans for individuals and enterprises, each valid for 3 months. Each lists 480P and 720P resolution options and clip lengths of 4 to 30 seconds, and says plan usage offsets per-token pricing, up to about 1:1.8 across resolutions and input modes. The snapshot showed "Loading pricing" where the dollar prices go, so no prices are quoted here.

Why is each plan a range?

The page says usage offsets per-token pricing, with up to a 1:1.8 spread across resolutions and input modes. A clip that uses reference video or a higher resolution consumes more of the plan than a plain text-to-video clip at 480p, so the low end of each range is the likelier match for a costly request and the high end for a cheap one. Plan on the low end if you use references.

The "valid for 3 months" line is the other half of the decision. A prepaid pack only pays off if you use it inside that window.

How does Sume price the same clip?

Sume does not sell a Seedance plan. The Video generation doc says billing is the workspace USD balance, reserved on submit at the provider list price times 1.25, and that usage.cost on the poll response is the Sume billable amount. The model's list basis is per 1,000 video tokens; read the live number from the catalog instead of a blog post.

Seedance 2.5 on Sume accepts 4 to 30 seconds at 480p, 720p and 1080p (Video Router). Note the plan page's resolution options stop at 720P; Sume's row adds 1080p.

import os, requests

r = requests.get(
    "https://api.sume.com/v1/video-router/models/seedance-2.5",
    headers={"Authorization": f"Bearer {os.environ['SUME_API_KEY']}"},
)
item = r.json()["data"]
print(item["capabilities"]["duration_seconds"], item["capabilities"]["resolutions"])
print(item["pricing"])

What should you ask before buying a plan?

Four questions settle it. How many 30-second clips will you really make in 3 months, at which resolution, and with how many reference videos? Do you need 1080p, which the plan page does not list? Can you use the whole pack inside the window? And how many clips must run at the same time, since that limit is set by the service you call, not by the pack?

If you cannot answer the first question with a number, a pay-per-request path carries less risk. If you can, compare the plan's seconds with a clip count you have already measured, using the low end of the range as the safe figure.

What does the plan page not tell you?

It does not show a fixed price in the snapshot, it gives seconds as approximate ranges, and it does not say what happens to unused seconds after the 3 months. Read those terms on the vendor's own page before you buy. On the Sume side the equivalent unknown is your clip count, which you can only measure by running a few real requests and reading usage.cost on each.

Which approach fits which use?

A prepaid plan suits a steady, predictable volume with a direct vendor account. Paying per request suits spiky or experimental use, such as a draft-and-final workflow where you do not know how many 1080p reruns you will need. In both cases the throughput limit is separate: on Sume, concurrency is plan-only and prepaid top-ups do not raise it (Generation admission). Check the live numbers on each side before you commit.