Google Chirp 3 vs Soniox STT-RT v5

Compare

Accuracy and speed

Left gets fewer words wrong. Lower has the final text ready sooner.

Accurate and fast0ms200ms400ms600ms800ms0%2%4%6%Words wrong on clean audio (Pipecat WER)
  • Chirp 32.00% · 475ms
  • STT-RT v52.54% · 38ms

Google Chirp 3 is ahead on both datasets together and the Pipecat Dataset, and Soniox STT-RT v5 on time to first text and final-text delay. The Ocular Dataset is too close to call.

Last updated Methodology by Dileep Chagam

Who is better at what

Chirp 3STT-RT v5
  • Overall WER2.76%3.02%
  • Pipecat Dataset WER2.00%2.54%
  • Ocular Dataset WER4.19%3.92%
  • Time to first text4.48s1.20s
  • Final-text delay475ms38ms

Lower is better on every measure. Too close to call means the likely ranges overlap. For the wider field, see our pick of the best voice AI APIs for real-time audio.

Clip by clip

On every clip both models were scored on, the lower WER wins, and WERs within 5 points of each other are a draw; in the rating, a wider win counts for more. Expected win rate is what the two ELO ratings predict for a match.

Chirp 3STT-RT v5
  • ELO rating979964–998 · rank 9–221000986–1007 · rank 6–20
  • Expected win rate47%53%
  • Ocular turns WER · 144 draws of 19728 won25 won
  • Pipecat WER · 835 draws of 999100 won64 won
  • Semantic WER · 177 draws of 1977 won13 won
  • Phone line · 152 draws of 19825 won21 won
  • Call-center noise · 114 draws of 19841 won43 won
  • Babble noise · 78 draws of 19856 won64 won
  • TV noise · 88 draws of 19892 won18 won
  • Reverb · 127 draws of 19828 won43 won
  • Short replies · 48 draws of 7013 won9 won
  • Entities · 46 draws of 502 won2 won

Short replies are decided on the error count and entities on exact capture, with no draw margin. How the rating is worked out

Chirp 3 vs STT-RT v5 FAQ

  • Google Chirp 3 is more accurate on both datasets. On the 1,000 public Pipecat clips, Google Chirp 3 scores 2.76% and Soniox STT-RT v5 3.02%, so Google Chirp 3 makes fewer errors by 0.26 points. On the 300 licensed Ocular turns, Google Chirp 3 scores 2.00% and Soniox STT-RT v5 2.54%, so Google Chirp 3 makes fewer errors by 0.54 points. Word error rate counts incorrect, missing, and extra words, so lower is better.

  • No. Soniox STT-RT v5 finishes the transcript 3.29s sooner. Google Chirp 3 has the final transcript 4.48s after the speaker stops and Soniox STT-RT v5 1.20s, at the median. Both are measured on the same 206 streamed turns.

  • It is a trade-off. Google Chirp 3 makes fewer errors on real conversations, while Soniox STT-RT v5 finishes the transcript 3.29s sooner. A voice agent cannot reply until the final transcript arrives, and the Ocular recordings are real conversations.

  • Yes, both return partial text before the speaker stops. Google Chirp 3 shows first text after 4.19% and Soniox STT-RT v5 after 3.92%, at the median.

Test your voice agent the same way.

Cekura runs these simulated calls on your agent and scores every one.