Sume API requests per minute by plan: write and read limits
Writes per minute are 120 on Free, 300 on Pro, 600 on Startup and 1200 on Scale. Reads are 40 times that, so 4800, 12000, 24000 and 48000.

Sume's API limits requests per minute by plan: writes are 120 (Free), 300 (Pro), 600 (Startup) and 1200 (Scale), and reads are 40 times writes. Enterprise is contract-based and behaves like Scale until limits are provisioned.
These are request limits, separate from how many generations run at once, which admission control sets.
The table
Numbers are from the Sume authentication page (read 2026-10-03). Prices of the plans themselves are on the API pricing page.
| Plan | Writes per minute | Reads per minute |
|---|---|---|
| Free | 120 | 4800 |
| Pro | 300 | 12000 |
| Startup | 600 | 24000 |
| Scale | 1200 | 48000 |
| Enterprise | Contract (Scale until provisioned) | Contract (Scale until provisioned) |
What counts as a read
Any GET or HEAD is a read, and so are /v1/generation/admission-preview and the MCP endpoint. Everything else, including every job submit, is a write. That makes polling cheap: one status check costs a read, and a read is a fortieth of a write.
The multiple comes from a deployment setting, so verify it from the headers instead of assuming it. A 429 carries retry-after and an error.details.scope of read or write.
curl -s -D - -o /dev/null https://api.sume.com/v1/me \
-H "Authorization: Bearer $SUME_API_KEY" \
| grep -iE "^(ratelimit-|retry-after)"Plan for the write limit
A burst of submits is limited by writes long before polling is limited by reads. At Free, 120 writes a minute is two per second, though the one running generation slot is the tighter limit in practice. See polling budgets by plan for the read side.
Sources
Related posts
More in Pricing
- Sume avatar quality defaults to plus when omitted: what it costs
Omit quality on an avatar video and Sume renders on Plus at $0.245 per second, 33 percent above Standard. How the default shows up in a 100-clip batch.
- Sume image markup at 10,000 images: list vs billed price per model
What the 1.25x house margin adds at 10,000 images per Sume image model, from $50 on Qwen Image and Grok Imagine to $527 on ChatGPT Image 2.
- Does Sume offer invoice or PO billing? What is public and what is not
The Sume pricing page lists invoice and PO billing under Enterprise with custom volume pricing. What is public, what is not, and how to budget before the call.
- Sume plans: Pro $40, Startup $120, Scale $400 and what they limit
What each Sume plan sets: monthly price, concurrent jobs, queue size and API write budget. Usage is billed at each model's rate, not by the plan.
Written by Sume