Sora's $0.10 to $0.70 per second: what replaces it
Luma's guide lists Sora at $0.10 to $0.70 per second, with the API sunset on Sept 24, 2026. What replaces it on Sume, which has no batch discount.

Luma's pricing guide says Sora cost $0.10 to $0.70 per second, with a Batch API at 50% off, and that the Sora API was sunset on Sept 24, 2026. The replacement on Sume is a per-second catalog id from GET /v1/videos/models; Sume does not offer a batch discount, so cut cost with lower resolutions and shorter drafts.
What the guide reports
These figures come from Luma's guide, read on the date shown. They describe a product that has been withdrawn, so use them as history to size a migration, not as a price you can still buy.
| Item | Reported |
|---|---|
| Price range | $0.10 to $0.70 per second |
| 720p to 1080p | 7x multiplier |
| Batch API | 50% off |
| API sunset | Sept 24, 2026 |
Pick a replacement id
Sume's catalog is queried, not guessed: GET /v1/videos/models returns each id with its resolutions, durations, aspect ratios and pricing_skus. Sume bills the provider list price times 1.25 on every model, reserved on submit from the workspace USD balance.
The docs name several ids with different limits: seedance-2.5 takes 4-30 seconds at 480p, 720p and 1080p, wan-3.0 takes 2-30 seconds, and gemini-omni-flash-1.1 takes 3-10 seconds with native audio. Other catalog models top out at 15 seconds. If you do not want to choose, model: "sume/auto" lets Sume pick, but pin an id when cost and output must repeat.
No batch discount: draft cheaper instead
The Sume docs reviewed for this post describe no batch tier or batch discount. The practical lever is resolution: models that offer 480p can be used for drafts, with the final render at 720p or 1080p only for the winners. Check the per-resolution price in pricing_skus first, since not every model prices resolution linearly.
Send an Idempotency-Key on every submit so a retried script does not double-bill.
- List the catalog before choosing an id.
- Draft at the lowest resolution the model offers.
- Re-run only winners at 1080p.
- Idempotency-Key on every submit.
# draft at 480p, then re-run the winner at 1080p
curl -X POST https://api.sume.com/v1/videos \
-H "Authorization: Bearer $SUME_API_KEY" \
-H "Content-Type: application/json" \
-H "Idempotency-Key: draft-001" \
-d '{
"model": "seedance-2.5",
"prompt": "A glass bottle on a stone ledge at sunrise, slow dolly in",
"resolution": "480p",
"duration": 6
}'Sources
Related posts
More in Pricing
- Sora back-catalog re-render: check balance before you submit
Sume reserves the estimated cost at submit and returns 402 insufficient_credits if it cannot. Read the balance, total the estimates, and size each wave to fit.
- Speech-to-text cost per hour: Scribe v2, MAI-Transcribe-2 and Sume
Per hour of audio: ElevenLabs Scribe v2 is $0.22, Scribe v2 Realtime $0.39, MAI-Transcribe-2-Streaming $0.54 through year end, and Sume STT $0.60.
- Seedance 2.5 clips: plan fee vs pay-per-use, break-even by volume
At full use Luma Plus is $1.17 per 5-second Seedance 2.5 clip and Pika Creator $1.36; Sume bills $2.89 a clip on top of its $40 Pro plan. Break-even math.
- Sume audio price sheet, October 2026: TTS, STT, detach, join, render
Every Sume audio step with its rate: TTS $0.0475 per 1,000 characters, STT $0.01 a minute, detach and join $0.01, render $0.10 a minute. Plus MAI prices.
Written by Sume