Eleven v4 or Sonic 3.6 for ad voiceover through an API

Pick on a listening test of your own scripts, not on a price table. On list price, Eleven v4 is cheaper per character ($0.022 per 1K characters, in a promo through Oct 12) than Cartesia Sonic at list ($0.038 per 1K). On reach, Eleven v4 claims 90+ languages and Sonic-3.6 claims 44. If you build on Sume, the choice is narrower: Sume serves Cartesia Sonic only, so Eleven v4 means calling ElevenLabs directly.

Both vendors publish blind-test claims (Eleven: about 75% listener preference; Cartesia: up to 93% in head-to-head tests). They are vendor-run with different baselines, so they cannot be compared with each other. Treat them as a reason to run your own test.

What each vendor states

What a 30-second ad read costs

Cartesia states that a minute of audio takes 750-800 credits, one credit per character. That rate is applied to both vendors as an assumption, because Eleven publishes no characters-per-minute figure on the pages read. A 30-second read is then 375-400 characters.

The Sume line is Sonic list x1.25 = $0.038 x 1.25 = $0.0475 per 1K characters. Sume rounds each job up to the cent, and a 5.5% platform fee applies to customer spend on top (Sume operations pricing policy, live since 2026-09-22).

Calling Sonic-3.6 through Sume

The TTS Router takes a required model (sonic-3.6, sonic-3.5, sonic-3, sonic-latest, sonic-preview), a transcript of 1 to 20,000 characters, and a voice selector. It is an async job: submit, then poll /v1/jobs/:id/status and read /v1/jobs/:id/result. The voice.id below is a placeholder for a voice you own.

curl -X POST https://api.sume.com/v1/tts-router/generate \
  -H "Authorization: Bearer $SUME_API_KEY" \
  -H "Content-Type: application/json" \
  -H "Idempotency-Key: ad-read-001" \
  -d '{
    "model": "sonic-3.6",
    "transcript": "Fresh coffee, ready in thirty seconds.",
    "voice": { "id": "YOUR_VOICE_UUID" },
    "language": "en"
  }'

How to decide

• Need audio tags, 90+ languages or voices beyond Sonic: use Eleven v4 from ElevenLabs. Sume does not ship it.
• Need one managed pipeline (voice, receipts, billing, avatars) with Sonic: use the Sume TTS Router.
• Either way, render the same three real scripts on both and have a native listener compare them before you commit a campaign.