CekuraBench

Inworld TTS-2

Every voiceCompare

The best of 26 voices at reading hard parts right

Hard parts read right: 93.4%

  • Also strong at
  • Reads numbers and shorthand rightTied best of 26 · 93%
  • Sounds natural2nd of 26 · 4.52

Inworld TTS-2: the best of 26 voices at reading hard parts right. Its weakest is medical terms, tied 12th of 26.

Last updated Methodology by Dileep Chagam

Inworld TTS-2 at a glance

Where it ranks among all 26 voices, in plain words.

Good at
  • Reads hard parts rightBest of 26
  • Reads numbers and shorthand rightTied best of 26
  • Sounds natural2nd of 26
  • Reads read from context rightTied best of 26
  • Starts talking fast3rd of 26
  • Reads codes and addresses rightTied 4th of 26
Not so good at

Nothing in the bottom quarter of the voices.

Hear Inworld TTS-2

The first read of each test sentence, unedited.

Inworld TTS-2

Inworld TTS-2 head to head

Pick any voice to set against it. Every pair has its own page.

Compare
MeasureInworld TTS-2Gradium TTS
  1. Non-streaming126 ms76 ms
  2. Streaming127 ms−64 ms
  3. StoppingNo stop messageNo stop message
  4. Accuracy93.4%79.5%
  5. Naturalness4.524.51
  6. Price / 1M$25.00$36.00
  7. Measures won3 of 62 of 6
The check marks the better value on each measure. Every pair has its own page with the full numbers.Open this pair · 325 pairs in all

Its closest rivals on speed

Inworld TTS-2 FAQ

  • Inworld TTS-2 starts in 126 ms on a typical sentence (median) and 148 ms on the slowest 1 in 10 (p90), 3rd of 26 voices.

  • Yes. With streamed text, Inworld TTS-2 starts 127 ms after the last word arrives.

  • Inworld TTS-2 has no stop message, so the agent has to close the connection to stop it.

  • Inworld TTS-2 read 93.4% of the hard parts right, as speech-to-text heard it. Its weakest kind is codes and addresses, at 85.7%. This is an estimate, not a listening test.

  • Inworld TTS-2 lists at $25.00 per million characters.