Start with 45 free minutes of callsTry for free

Text-to-Speech · Multilingual

Korean

Language

Request a new language

Tell us which language you want added to this benchmark.

Benchmarks

X
Y

Naturalness

higher is better

1500 is this language's own field — not comparable across languages.

field averagetied with bestAPAWS PollygenerativeTry →953RimeRimecodaTry →1187MiniMaxMiniMaxspeech-2.6-hdTry →1407AzureAzureAI Speech (native)1438OpenAIgpt-4o-mini-ttsTry →1498ElevenLabseleven_v3_conversationalTry →1628InworldInworldtts-2Try →1722Cartesiasonic-3.5Try →1859GeminiGemini3.1-flash-tts-preview1917
400800120016002000

Time to first audio

lower is better

The pause the caller hears before the agent starts speaking.

InworldInworldtts-2Try →122ms · 134ms
Cartesiasonic-3.5Try →132ms · 246ms
RimeRimecodaTry →223ms · 253ms
APAWS PollygenerativeTry →241ms · 278ms
MiniMaxMiniMaxspeech-2.6-hdTry →336ms · 373ms
OpenAIgpt-4o-mini-ttsTry →449ms · 855ms
GeminiGemini3.1-flash-tts-preview678ms · 1.02s
ElevenLabseleven_v3_conversationalTry →774ms · 939ms
0ms300ms600ms900ms

p50 p90

Robustness

higher is better

Whether the voice says the right words for a number, date, currency amount or operator.

Gemini0.75
InworldTry →0.60
Try →0.50
RimeTry →0.00
Try →0.00
APTry →0.00
Try →0.00
MiniMaxTry →0.00
0.000.250.500.751.00

Cost

lower is better

$ per 1M characters · vendor list price, the same in every language