Face swap beta on a 13-second clip at plus: reserves $3.68, not $3.19

Sume Avatar Face Swap (beta) reserves the 15-second price for any source clip: 15 x $0.245 = $3.675 at plus. A 13-second clip at that rate would be $3.185.

5 min readSume
All posts

The catalog reserves $3.675 for a plus-tier Avatar Face Swap job, and it shows the same figure as minimum and maximum, as of 2026-10-09. A 13-second source clip priced at the plus rate of $0.245 per second would be 13 x 0.245 = $3.185, so the catalog figure is $0.49 higher. The pricing basis in the catalog says the beta reserves the Avatar Video public price for the maximum beta source-video duration, and that final workflow pricing may be refined before general availability.

Treat $3.675 as the number to hold in your balance for any plus swap, and expect it to change before GA.

The three reservations

Face swap uses the same per-second rates as avatar video: $0.184 standard, $0.245 plus and $0.55 max. The beta multiplies each by 15 seconds, the longest source clip it plans for.

Face swap beta reservation by tier (catalog, read 2026-10-09)
TierRate per second15-second reservation13-second clip at the same rate
standard$0.184$2.76$2.392
plus$0.245$3.675$3.185
max$0.55$8.25$7.15

What the docs require of the source clip

The endpoint is POST /v1/models/sume/avatar-face-swap/v1.0/runs. It needs avatar_handle, video_url and quality. There is no default for quality: if you leave it out, the request does not fall back to plus. The source video must be a public HTTPS URL, and the endpoint rejects localhost, private-network, signed or private URLs, and provider task URLs. The worker plans for roughly 4-15 seconds of source with usable audio.

The endpoint does not take prompts, transcripts, duration knobs, aspect ratio or provider fields. If you want the avatar to say new words, use script-driven avatar video instead.

Choosing between a swap and a scripted video

On price alone, a 13-second scripted avatar video at plus is 13 x 0.245 = $3.185, which is lower than the $3.675 swap reservation. Face swap earns its place when the performance in the source clip is the point: timing, gesture and delivery that you cannot get from a script. If the words are what matter, script the video. If the original take matters, swap the face and keep the take.

Either way, poll the job at GET /v1/jobs/:id/status and read resource_status for readiness, as the face-swap page advises. See the beta limits post for the clip-length rules.

Budgeting a batch

If you plan ten swaps at plus, hold 10 x $3.675 = $36.75, not 10 x $3.185 = $31.85. The $4.90 gap is the cost of the beta's fixed reservation. At standard the same ten swaps reserve $27.60, and at max $82.50.

Because the catalog may be refined before GA, re-read the live catalog at https://api.sume.com/v1/catalog before each large run instead of caching these numbers. The face-swap page marks the endpoint as Beta, and a Beta endpoint is the one place on this list where pricing and limits are most likely to move.

Sources

Related posts

More in Pricing

All Pricing posts

Written by Sume