What a $10 Sume wallet top-up buys, job by job at fixed prices

The smallest wallet purchase is $10. At fixed Sume prices that is 50 caption jobs, 80 music generations, 1,000 transcribed minutes or 210,526 TTS characters.

5 min readSume
All posts

The Sume pricing FAQ says teams can fund a shared wallet with on-demand purchases from $10 to $1000, and generation and API calls are billed at each model's published USD rate. At the fixed per-unit rates in the repo, $10 buys 50 caption jobs, 80 music generations, 1,000 minutes of transcription, 100 output minutes of timeline render or 210,526 characters of text to speech, but not all at once. Read 2026-10-03 from the Sume pricing page and repo.

This post is the division, laid out so you can plan a first top-up. It uses only fixed-price jobs; video generation is priced per second by model and has its own tables.

The table

Each row divides $10 by the published rate and rounds down to a whole unit. Fixed rates are the public prices in the repo; video captions and Music 1.0 are also stated in their docs.

What $10 buys at Sume's fixed prices (read 2026-10-03)
JobRateUnits for $10
Standalone video caption job$0.20 per job (up to 60 s)50 jobs
Music 1.0 generation$0.125 per generation80 generations
Text to speech$0.0475 per 1,000 characters210,526 characters
Speech to text$0.01 per minute1000 minutes
Timeline render$0.10 per output minute100 output minutes
Background removal$0.0225 per image444 images
Image upscale$0.20 per image50 images
Video upscale$0.009 per second1,111 seconds

A worked mixed budget

A small ad test might use 10 caption jobs ($2.00), 8 music generations ($1.00), a 2,000-character voiceover ($0.095) and 5 minutes of timeline render ($0.50). That totals $3.595, so the first $10 covers it with $6.405 left over. The numbers are list arithmetic on the fixed prices, not a measured bill.

A few small notes: a caption job reserves $0.20 whatever the clip length up to 60 seconds, so a 10-second clip costs the same as a 55-second one.

What a top-up does not cover

Video generation is billed per second and varies by model and resolution, so a $10 balance buys different amounts depending on the route. Wan 3.0 at 480p is $0.05 list, billed at list times 1.25, so $10 is 160 seconds at that tier. Avatar video is $0.184 a second at Standard, so $10 is 54 seconds. The two rows are the extremes of what a top-up buys in moving picture.

The plan grid is separate. Plans price concurrency and access, and usage is billed at published rates; the FAQ does not describe the subscription fee as a credit toward usage, so budget the two separately.

Limits

Start with one job of each kind you plan to run and read the wallet after.

  • The $10 minimum is from the pricing FAQ and may change; check the page before you rely on it.
  • Units are rounded down to whole units in the table.
  • Video rates change by model; see the Video Router docs.

Sources

Related posts

More in Pricing

All Pricing posts

Written by Sume