AI voiceover cost for 100 scripts: Sume vs ElevenLabs v4

For 100 voiceover scripts of 800 characters each (80,000 characters, roughly one minute of speech each), Sume TTS 1.0 costs $3.80. ElevenLabs lists v4 at $0.08 per 1,000 characters, which is $6.40, and a 72 percent introductory price of $0.022 that ends October 12, which is $1.76 (all read 2026-10-07). The cheapest option changes with the model and the plan you are on.

How the numbers are built

Sume prices TTS per transcript character at $0.0475 per 1,000, which the docs describe as the Cartesia list times 1.25 (so the implied list is $0.038). Spaces and punctuation count, and one request takes up to 20,000 characters. Cartesia's page says one minute of audio takes 750 to 800 credits at one credit per character, so 800 characters is the upper end of a minute.

The formula is price per 1,000 x 80. Sume: $0.0475 x 80 = $3.80. ElevenLabs v4: $0.08 x 80 = $6.40. Discounted: $0.022 x 80 = $1.76.

What the table hides

The Cartesia rows are subscription-derived: the price is the plan fee divided by included credits, so you only get that rate if you actually use all the credits. The ElevenLabs discount is time-limited. The fal row is ElevenLabs Multilingual v2 resold on fal at $0.10, which is above ElevenLabs' own $0.04 for v2.

Voices are not equivalent. Sume TTS takes a voice selector (an avatar or a TTS voice id), and its docs say ElevenLabs and OpenAI engines are out of scope for the router's first ship. The comparison is price per character, not voice quality.

When Sume still makes sense

A voiceover is rarely the biggest line. A 15-second voiceover is about 200 characters, so $0.0095 on Sume, against several dollars for 15 seconds of generated video. Keeping voice, video and captions on one balance and one job model is the draw. If you only need voice at scale, a subscription from a voice vendor will undercut a pay-per-character rate.

Scaling to 1,000 scripts

At 800 characters, 1,000 scripts is 800,000 characters. Sume: 800 x $0.0475 = $38.00. ElevenLabs v4 at list: 800 x $0.08 = $64.00. ElevenLabs v2 or Flash: 800 x $0.04 = $32.00. Cartesia Scale: 800 x $0.037375 = $29.90. The ranking holds at every volume because every figure is linear in characters.

What changes at volume is whether a subscription is worth it. A Cartesia Scale plan has 8,000,000 credits, enough for 10,000 such scripts, so 1,000 scripts uses 10 percent of it. If you use only that, you are paying $299 for $29.90 of speech. Pay-per-use rates win at low volume; plans win only when you fill them.