A 37-second AI avatar video: six Sume prices from $6.81 to $21.46

A 37-second talking avatar video on Sume costs $6.808 at the standard tier, $9.065 at plus and $20.35 at max, when you send no product image. With a product image the same 37 seconds cost $7.178, $9.546 and $21.46. The price is per second, so the arithmetic is one multiplication.

Thirty-seven seconds is an awkward length on purpose. It sits inside the 4 to 60 second window that Sume accepts for one avatar video job, so one request covers it, and it does not land on a round number that makes the per-second rate easy to forget.

The six prices, worked

The rates come from the Sume catalog (read 2026-10-09): standard is $0.184 per second without a product image and $0.194 with one, plus is $0.245 and $0.258, and max is $0.55 and $0.58. Multiply each by 37.

What the catalog reserves at submit

The catalog says a submit-time reservation uses exact USD micros and rounds up to cents for balance compatibility, then captures on completion and refunds a failed or pre-generation canceled job. So the plus tier without a product image is $9.065 in the arithmetic, and the balance check works against the rounded-up figure of $9.07. The standard tier reserves $6.81 and max reserves $20.35.

If the balance is short, the request is refused with 402 insufficient_credits, as the errors page lists. Nothing is generated and nothing is billed.

• Standard, no product image: $6.808 (reserve $6.81)
• Plus, no product image: $9.065 (reserve $9.07)
• Max with a product image: $21.46 (reserve $21.46)

Pick the tier before you pay for the render

The docs describe plus as the default when you omit quality, standard as the fastest Sume execution path and max as the highest quality tier with slower turnaround. A cheap way to choose is to run an avatar video preview first. Preview stills are tier-independent, so you can approve the first frame and then set the final quality on the generate-video call without making a new preview.

The difference between tiers on 37 seconds is large: max costs $13.542 more than standard without a product image. If the clip is a draft for review, standard is the sensible start.

curl -X POST https://api.sume.com/v1/avatar-1.0/talking-video \
  -H "Authorization: Bearer $SUME_API_KEY" \
  -H "Content-Type: application/json" \
  -H "Idempotency-Key: avatar-37s-001" \
  -d '{"avatar_handle": "product_host", "quality": "standard", "script": "<your 37-second script>"}'

Limits to remember

The script must estimate to 4-60 seconds inclusive, and resolution is 720p at this time. Aspect ratios are 1:1, 3:4, 9:16, 4:3 and 16:9, with 9:16 as the default. Poll the job at /v1/jobs/{id}/status and fetch the result when it is completed; see the jobs page for the status table.