How much does a 40-second AI video cost on Sume, by model and split?
A 40-second sequence costs $2.51 on Wan 3.0 at 480p, $5.00 at 720p and $23.12 for Seedance 2.5 at 720p. Prices for each split into legal-length jobs.
No Sume model makes 40 seconds in one job, so you split it. At 720p, 40 seconds costs $5.00 on Wan 3.0 (30 + 10), $23.12 on Seedance 2.5 and $5.00 on Gemini Omni Flash 1.1 (four 10-second jobs). Wan at 480p is $2.51.
Price by model and split
The longest clip is 30 seconds (seedance-2.5, wan-3.0) or 10 (Omni), per the Sume video models docs. Price is linear in seconds, so any split costs the same apart from cent rounding.
| Model | Split | 480p | 720p | 1080p |
|---|---|---|---|---|
| wan-3.0 | 30 + 10 s | $2.51 | $5.00 | $10.00 |
| seedance-2.5 | 30 + 10 s | $10.76 | $23.12 | $56.87 |
| gemini-omni-flash-1.1 | 4 x 10 s | $1.52 at 360p | $5.00 | $7.52 |
Splitting trade-offs
Cutting at a scene boundary matters more than saving a cent. Each job is independent, so the seam is yours to hide with a cut or a fade. A 40-second Omni sequence costs the same as one Wan at 720p, but takes four jobs and four seams.
If the script is a single scene, use the longest clip so there are fewer seams. The step-by-step method for 30 + 10 on Sume, with the timeline join, is in the 30 + 10 post.
- Wan 480p: $2.51.
- Omni and Wan at 720p: $5.00.
- Seedance 2.5 at 720p: $23.12.
Planning the seam
With two jobs, the seam falls at a point you choose, so make it a natural cut: a scene change, a title card or a beat in the music. If the final scene must match the last frame of the first, that is an input-reference decision for the second job, and on Seedance 2.5 a reference video changes the price, so estimate the second job with its references included.
For four Omni jobs, three seams is a lot. Consider whether a 30 + 10 split on Wan at the same $5.00 gives you a better edit, or whether 20 + 20 is needed for pacing (not possible on Omni, whose cap is 10 seconds). The price is the same either way, so decide on the cut and pay accordingly.
How these numbers were produced
I read the rate tables in the clone of Sume's repository on 2026-10-06 and called the three estimators for those models for wan-3.0, seedance-2.5 and gemini-omni-flash-1.1, covering the 30 + 10 and 4 x 10 second splits. The output fields are billableAmountUsdMicros and billableAmountUsdCents; the tables show the cents.
These are estimates of the billed amount for the stated settings. Anything not in the table, such as a different frame shape, a reference input or a changed quality tier, can change the number, so run the estimate for your own request before a large batch. Where this page and the public API pricing list disagree, the pricing page and the live quote win.
Sources
Related posts
More in Pricing
- 4K AI video API: which Sume models accept 4K and what a clip costs
Only Gemini Omni Flash 1.1 and MiniMax H3 accept 4K in Sume's Video Router. A 10-second Omni clip is $3.75; a 5-second H3 upscale is $1.00.
- Sume's 5.5% fee in dollars on a $50, $250, $1,000 or $5,000 month
The 5.5% fee is charged on the list x 1.25 amount. On $50 of catalog spend it is $2.75, on $1,000 it is $55. A table, and where the fee sits in the formula.
- 500 flashcard phrases on Sume TTS: $5.00 per job, $1.43 batched
500 phrases of 60 characters: one job each bills $5.00 because each rounds up to a cent; batches of 100 bill $1.45 and two capped jobs $1.43.
- Agent 365 cost management for Copilot Studio agents vs a per-run cap
Agent 365 sets spend policy per user group and adds Copilot Studio agents in October. A per-run cap on the API call bounds the one task. Know which you need.
Written by Sume