ElevenLabs v4 voice cost vs a 30-second avatar clip: under 1%

On ElevenLabs' API pricing page, the v4 model lists $0.08 per 1,000 characters at the regular price and $0.022 at a 72 percent discount that runs until October 12. A 30-second avatar script is roughly 450 characters, so the voice is about 3.6 cents at the regular price, or 1 cent discounted. A 30-second Sume avatar clip on plus is $7.35, so voice is under half a percent of the bill.

The arithmetic

I assumed 150 spoken words a minute and about six characters per word, so 75 words, about 450 characters, for 30 seconds. Change those inputs and the result scales, but the proportion barely moves.

What the avatar costs

The Sume rate is per second of finished video, $0.184, $0.245 or $0.55 by tier. For 30 seconds that is $5.52, $7.35 or $16.50. The avatar video guide takes a script and an avatar handle, and the catalog lists one per-second rate, not a separate voice line.

So when you compare tools, compare the face. Saving a cent on the voice does not change the total, but moving from standard to max changes it by $10.98 on a 30-second clip.

Where voice price does matter

Voice price matters at the audio-only scale: long narrations, audiobooks, thousands of short prompts. It does not matter when each voice line also needs a rendered face. If you want a specific outside voice on a face, Sume has a separate still-plus-audio route, listed in the models overview. Read its audio rules before you plan around it, and price it separately, because it bills by audio second rather than through the script route.

The discount deadline

The v4 and v4 Turbo discounts on the page run until October 12. After that the regular prices apply: $0.08 and $0.04 per 1,000 characters. For a script of a few hundred characters the difference is a few cents, so a deadline should not drive an avatar budget.