Kokoro
"Kokoro is an open-weight TTS model with 82 million parameters." Despite its lightweight architecture it "delivers comparable quality to larger models while being significantly faster and more cost-efficient".
text → audio · made by Hexgrad
$0.000001per character
DeepInfra · 2 sellers
Sold by 3 ways
| Seller | Lane | Rate |
|---|---|---|
| DeepInfra aggregator | standard | $1per million charactersapi.deepinfra.com · read 2026-08-25 |
| Together AI 0 ways | $0.000004 | |
| Together AI aggregatornot on the seller's list today | standard | $4per million charactersdocs.together.ai · read 2026-08-24 |
| Featherless AI aggregator | standard | $0.3per Mtok in$1.2per Mtok outapi.featherless.ai · read 2026-09-20 |
Measured 1 standing
| Place | Board | Metric | Score |
|---|---|---|---|
| 32ndof 42 | TTS Arena V2 | Elo | 1464 |
About
Kokoro — an audio model from Hexgrad, sold by 2 companies from $0.3 in and $1.2 out per million tokens, placed 32nd of 42 on TTS Arena V2.
It takes text and returns audio. The catalogue files it under speak. Two companies sell it. The cheapest is $0.3 in and $1.2 out per million tokens at Featherless AI. It stands 32nd of 42 on TTS Arena V2.
Every current figure
- Maker
- Hexgrad
- Register
- model
- Takes
- text
- Returns
- audio
- Licence
- not read
- Sellers
- 2
- Maker's own price
- not read
- Price
- $0.3 in and $1.2 out per million tokens — Featherless AI
- Boards
- 1
- Best place
- 32nd of 42 — TTS Arena V2
Known as 2 names
Kokoro v1.0hexgrad/Kokoro-82M