Seedance Shot Budget: Mini and Fast in 1,500 Credits

A 20-second edit can require 60 generated seconds

For an editor preparing a community sports event teaser, four five-second shots sound like a twenty-second purchase. That is only the footage selected for the edit. If each shot receives two draft attempts and one fresh final attempt, the generation plan contains sixty seconds. Treating the final running time as the billable quantity hides two thirds of the planned work before the first request is sent.

With the Orux rates checked on October 30, 2026, that exact plan totals 1,069.8 Credits with seedance-2.0-mini or 1,720.968 Credits with seedance-2.0-fast. Drafts use 480p and finals use 720p. Against a hypothetical allowance of 1,500 Credits, the first plan leaves 430.2 Credits; the second exceeds the allowance by 220.968 Credits. These are deterministic planning calculations, not results from generating or judging videos.

This guide gives you a reusable shot ledger, a price comparison with explicit conditions, and runnable Python that checks the whole plan before you purchase anything. Its central decision is how much generation to reserve, rather than which model produces the best footage. No paid model call was made for this article.

Define the deliverable before comparing models

Use this fictional brief: a twenty-second community race teaser, made from four independently generated five-second shots. The editor needs an opening location, participant preparation, an action moment, and a closing event card. These are planning labels, not prompts whose execution has been verified. Any exact event text can be added in the editing application rather than assumed to appear correctly inside generated footage.

For each shot, reserve two 480p drafts for composition and motion discussion, followed by one new 720p generation for the final selection. A draft is a separate paid output in this plan. A final is another new generation; the calculation assumes neither a free upgrade nor that the new generation reproduces an approved draft exactly. Continuity, readable details and usable motion still need review after an actual result exists.

Both current Orux forms expose 480p and 720p choices, and integer durations from four to fifteen seconds. Five seconds therefore fits the checked form constraints. A three-second insert or a sixteen-second shot does not fit this calculator. Change the editorial plan instead of silently rounding an unsupported duration into a purchase.

The forms also distinguish first/last-frame inputs from a separate reference-image mode. The budget below is not proof that every reference combination works. Before using references, verify the current form, rights to the assets and the estimate for your exact request. Vendor descriptions of the Seedance family should not be read as a promise that every native feature is exposed by these two Orux entries.

Compare the exact resolution row

Swipe horizontally to view the full table.

Rates are Credits per generated second. One Credit equals $0.005. The table comes from the public per-second price rows for the named aliases and resolutions, not from a headline “from” rate. At the checked time, those rows specify exact quantities and no minimum charge. A future configuration or a differently conditioned request can change the estimate, so refresh the model page before committing funds.

We did not verify an equivalent first-party Mini/Fast price schedule for these exact conditions. This article therefore makes no official-price savings claim and no claim that Orux beats a vendor or competitor. Likewise, “fast” in an alias is not a measured turnaround time: the calculations do not rank latency, visual quality or acceptance rate. Those would require separate evidence and a comparable task.

A five-second Mini shot at 720p is 141.59 Credits, while a five-second Fast shot at 720p is 227.774 Credits. These amounts are useful retake units. They should not be confused with the cost of one finished five-second segment, because a rejected attempt still belongs in the production ledger.

Build a ledger with separate draft and final quantities

Swipe horizontally to view the full table.

Read the rows as four production reservations. The four opening-to-closing labels identify where a decision belongs; each row has the same arithmetic. Together they reserve forty draft seconds and twenty final seconds. A single five-second clip containing multiple internal camera changes would be a different creative request, even if its length matched one row. This guide does not claim shot-by-shot timing control inside such an output.

For Mini, the draft reserve is 40 × 12.586 = 503.44 Credits. The final reserve is 20 × 28.318 = 566.36 Credits. Adding them yields 1,069.8 Credits. For Fast, the same quantities are 809.872 and 911.096 Credits, totaling 1,720.968 Credits. Keep the resolution with each quantity; multiplying all sixty seconds by one rate would describe a different plan.

The 1,500-Credit allowance is an invented project constraint, not a required deposit, platform minimum or recommended spending level. Choose your own cap. Nor is two drafts per shot an observed success rate. It is a reservation the editor can increase, reduce or decline after discussing the actual brief. If the first output works, unused allowance remains unused; a plan does not require spending every reserved Credit.

Run the calculator without a model request

Save the following as budget.py and run python 3 budget.py. It uses only Python’s standard library, makes no network request and does not read an API key. Rates are held as integer micro-units internally, with decimal conversion only for display. Readers see and approve Credits; the internal representation keeps repeated multiplication exact.

"""Offline planning example. Rates checked 2026-10-30; refresh before purchase."""
from decimal import Decimal

RATES = {
    ("seedance-2.0-mini", "480p"): 12_586_000,
    ("seedance-2.0-mini", "720p"): 28_318_000,
    ("seedance-2.0-fast", "480p"): 20_246_800,
    ("seedance-2.0-fast", "720p"): 45_554_800,
}

def cost(model, resolution, seconds, attempts=1):
    if type(seconds) is not int or not 4 <= seconds <= 15:
        raise ValueError("Duration must be an integer from 4 to 15 seconds")
    if type(attempts) is not int or attempts < 1:
        raise ValueError("Attempts must be a positive integer")
    return RATES[(model, resolution)] * seconds * attempts

def credits(micro):
    return format(Decimal(micro) / Decimal(1_000_000), "f")

def plan(model, draft_resolution, final_resolution, draft_attempts):
    # Four separate five-second shots; final pass is a fresh generation.
    draft = 4 * cost(model, draft_resolution, 5, draft_attempts)
    final = 4 * cost(model, final_resolution, 5)
    return draft + final

if __name__ == "__main__":
    cap = 1_500_000_000  # Hypothetical project allowance: 1,500 Credits
    for model in ("seedance-2.0-mini", "seedance-2.0-fast"):
        total = plan(model, "480p", "720p", 2)
        remaining = cap - total
        # Never report negative remaining funds as usable retake capacity.
        extra = max(0, remaining // cost(model, "720p", 5))
        print(model, credits(total), credits(remaining), "extra_5s=", extra)

The saved run printed:

seedance-2.0-mini 1069.8 430.2 extra_5s= 3
seedance-2.0-fast 1720.968 -220.968 extra_5s= 0

The third output column (counting the model alias first) is remaining allowance, which may be negative. Extra five-second retakes are counted only when remaining allowance is positive. The program rejects non-integer durations, durations outside four to fifteen seconds, and invalid attempt counts. A missing model/resolution combination raises an error rather than quietly substituting another price.

Local checks confirmed both plan totals and rejected durations of 3, 16 and 5.5, plus a Boolean masquerading as an integer. This is an arithmetic and input-validation exercise. It cannot tell whether a generated athlete, camera movement or event logo is usable. Keep that distinction when adapting the script into a spreadsheet or production tool.

Spend headroom only after reviewing the actual shots

The Mini plan leaves 430.2 Credits. At the checked 720p rate, that covers three additional five-second attempts: 3 × 141.59 = 424.77 Credits, leaving 5.43 Credits. A fourth such attempt would not fit. This is purchasing capacity under the hypothetical cap, not a promise of three acceptable shots.

The Fast plan starts 220.968 Credits over the same cap, so its extra-retake capacity is zero. Do not treat a negative remainder as a small inconvenience to be silently billed. Revise the plan before generation. For example, one 480p draft per shot followed by one fresh 720p final per shot would total 1,316.032 Credits for Fast. It fits numerically, but reserves less opportunity for draft exploration; the team must accept that trade-off.

Do not mix model rates inside a single assumed output. If you deliberately draft with one alias and generate finals with another, enter that as a separate plan and review continuity risk explicitly. A change in model or resolution can change the result’s appearance; a budget calculator cannot preserve a person’s identity, garment or scene geometry. Saving money on a line item does not establish equivalence of the deliverable.

Keep actual task costs beside planned costs once you begin. Mark each result accepted, rejected or awaiting review, and record the reason. Only actual billing and accepted outputs can establish the eventual cost per usable second. Until then, a reservation is useful precisely because it exposes what remains uncertain rather than pretending the edit is already finished.

Recheck the live request before starting production

Open seedance-2.0-mini and seedance-2.0-fast, select the intended resolution and duration, and compare the current estimate with the ledger. If a value differs, stop and update the rates before comparing plans. The script’s dated constants are a reproducible example, not a permanent catalog.

Bring a defined deliverable to review: four shots, their intended editorial roles, draft count, final count, selected resolution and maximum authorized Credits. Agree on who can approve additional attempts. That produces a useful production conversation without asserting that a preset number of generations will solve the creative task.

For judging actual footage, the existing Kling two-shot review shows why result selection is a separate activity from request planning. If your project also includes retrieval-based text work, the Qwen3 RAG budget guide uses different metered quantities; do not carry a per-second rate into an embedding calculation.

FAQ

Q.Does a twenty-second edit mean twenty billed seconds?

Only if you generate exactly twenty seconds once and use everything. In the worked plan, four shots each receive two draft attempts plus one fresh final attempt, giving sixty generated seconds. The final timeline remains twenty seconds. Rejected attempts and unused variations still belong in your cost ledger.

Q.Is a final generation a free upgrade from a draft?

No free upgrade is assumed here. The plan budgets each final as a new 720p generation after separate 480p drafts. It does not establish that a saved draft can be upgraded or reproduced exactly. Check the current model form and estimate before relying on any different workflow or feature.

Q.Can 1,500 Credits guarantee four usable final shots?

No. The allowance constrains planned purchases; it does not determine acceptance of the outputs. The Mini example leaves capacity for three extra five-second 720p attempts, but none is guaranteed to meet the brief. Review actual results and obtain approval before extending a plan beyond the agreed spending cap.

Q.Does the calculator measure Mini versus Fast quality or speed?

No. It compares dated Orux price rows under explicitly matched duration and resolution choices. No videos were generated, timed or scored for this article. Model naming and vendor-family descriptions do not substitute for a controlled comparison. Use a separate, authorized evaluation if quality or turnaround is the deciding factor.

Q.How should I refresh this budget for another project?

Recheck the live aliases, supported durations, resolution conditions and current estimate, then replace the rate constants and shot quantities. Run the local checks again and choose your own cap. Preserve separate draft and final rows so a changed number cannot hide additional generations or imply that one result automatically becomes another.

Prices checked on 2026-10-30. Sources: Mini public catalog, Fast public catalog, and ByteDance Seedance family information. The first two sources establish Orux options and prices; the family page supplies context only, not an exact Mini/Fast tariff.