Tied for best of 28 voices at reading read from context right
Read from context: 100%
Grok TTS: tied for best of 28 voices at reading read from context right. Its weakest is naturalness, 27th of 28.
Last updated Methodology by Dileep Chagam
Where it ranks among all 28 voices, in plain words.
The first read of each test sentence, unedited.
Pick any voice to set against it. Every pair has its own page.
Grok TTS starts in 358 ms on a typical sentence (median) and 446 ms on the slowest 1 in 10 (p90), 19th of 28 voices.
Grok TTS starts in 358 ms on a typical sentence (median) and 446 ms on the slowest 1 in 10 (p90), 19th of 28 voices.
Yes. With streamed text, Grok TTS starts 157 ms after the last word arrives.
Yes. After the stop it still sends 0 ms of audio (median), which the agent throws away.
Grok TTS read 87.4% of the hard parts right, as speech-to-text heard it. Its weakest kind is numbers and shorthand, at 82.1%. This is an estimate, not a listening test.
Grok TTS lists at $15.00 per million characters.
Cekura runs these simulated calls on your agent and scores every one.