whatmodel.sh / tts / Sonic 3
Sonic 3
↗ share
+ contribute
Sonic 3 is a closed-weights text-to-speech model from Cartesia, released October 2025. Licence: proprietary.
Catalogue identity notes
Vapi registry cartesia-sonic-3, apiModelId sonic-3. Distinct from Sonic 3.5 and Sonic 3.6. Rating-only comparison identity; no price, API or run metadata asserted.
Score within the Vapi Humanness · Bradley–Terry rating, 2026-09-25 snapshot comparison set only.
$ how it compares
The text-to-speech models ranked either side of Sonic 3, out of 148 with a score, then more from Cartesia.
| # | model | score | where | ||
|---|---|---|---|---|---|
| 55 |  CosyVoice v3 Plus | 58±4 | api $26.00/M chars | ||
| 56 |  Octave TTS (Octave 1) | 58±4 | api $50.00/M chars | ||
| 57 | Sonic 3 | 58±2 | api $65.00/M chars | ||
| 58 |  GPT-4o mini TTS | 58±2 | comparison only | ||
| 59 | Sonic 2 | 58±3 | comparison only | ||
| more from Cartesia | |||||
| 35 | Sonic 3.5 | 63±2 | comparison only | ||
| 105 | Sonic (1) | 42±3 | comparison only | ||
see all text-to-speech models · find the ones that fit your machine
$ what it's good at
other tests5 not in a task · 98% of its score
| test | result | vs the others in its run | counts |
|---|---|---|---|
| Vapi Humanness · Bradley–Terry rating, 2026-09-25 snapshot | 1185.7 ✎ | ▁···▁▃··█▄▁▇ beats 5/18 | 47% |
| Vapi Humanness · historical Elo, 2026-06-10 | 1096.7 ✎ | ▃·▃▃▃·▃██▃·▃ beats 1/12 | 24% |
| Rime ContentBench · paired human preference vs Rime Coda (US English) | -0.20 / 2 ✎ | ▄·▄········█ beats 1/3 | 11% |
| Rime SupportBench · paired human preference vs Rime Coda (US English) | -0.20 / 2 ✎ | █·······█·██ beats 2/3 | 10% |
| Voice-First Nation · 10 Indian languages overall pairwise preference (Bradley–Terry) | 1050.8 ✎ | ▄····▄·····█ beats 2/3 | 6.9% |
$ where to run it
Through an API
api · Together AI · $65.00/M chars · www.together.ai ↗ · price sourceUSD per 1M input characters: Together's serverless model list gives Cartesia Sonic 3, API model cartesia/sonic-3, at $65.00 per 1M chars (the Together pricing page also lists 'Cartesia Sonic-3 $65.00' per 1M characters). Cartesia's own pricing page prices credits and names only Sonic-3.6, so Together's per-model rate is used. Product: Together serverless text-to-speech. API model: cartesia/sonic-3. Settings: standard. Price basis: published list rate. Currency: USD. Together official docs, observed 2026-09-25.
$ how the score was worked out · The Humanness Index by Vapi ·… 47%, The Humanness Index by Vapi ·… 24%, Rime Alpha TTS Benchmark v1.0… 20% and 1 more
Each result implies a score on its own, after adjusting for how hard its test is and who ran it. The score is their weighted average, with a small pull toward the middle for thinly tested models. Each test's weight is its counts column above; here it's added up by who ran the tests. How it works
| source | results | share of the score |
|---|---|---|
 The Humanness Index by Vapi · September 2026 public standings | 1 | 47% |
| The Humanness Index by Vapi · June 2026 production export | 1 | 24% |
 Rime Alpha TTS Benchmark v1.0.0 · SupportBench and ContentBench paired human preference | 2 | 20% |
 Preferences of a Voice-First Nation: pairwise human evaluation of TTS in 10 Indian languages (arXiv:2604.21481v2) | 1 | 7% |
| prior · pull toward the middle | 2% | |
| weighted average | = 57.7 |
$ improve this page
Use a short GitHub form or edit the data through a pull request. Every number needs a source; here's which ones we can use.
data updated · sources and licences · json