ElevenLabs Scribe v3 Live vs Soniox STT-RT v5
Accuracy and speed
Left gets fewer words wrong. Lower has the final text ready sooner.
- Scribe v3 Live1.81% · 57ms
- STT-RT v52.54% · 38ms
ElevenLabs Scribe v3 Live is ahead on both datasets together, the Pipecat Dataset, the Ocular Dataset and time to first text, and Soniox STT-RT v5 on final-text delay.
Last updated Methodology by Dileep Chagam
Who is better at what
- Overall WER1.89%3.02%
- Pipecat Dataset WER1.81%2.54%
- Ocular Dataset WER2.05%3.92%
- Time to first text685ms1.20s
- Final-text delay57ms38ms
Lower is better on every measure. Too close to call means the likely ranges overlap. For more on the platform, read our ElevenLabs review.
Scribe v3 Live vs STT-RT v5 FAQ
Which is more accurate, ElevenLabs Scribe v3 Live or Soniox STT-RT v5?
ElevenLabs Scribe v3 Live is more accurate on both datasets. On the 1,000 public Pipecat clips, ElevenLabs Scribe v3 Live scores 1.89% and Soniox STT-RT v5 3.02%, so ElevenLabs Scribe v3 Live makes fewer errors by 1.13 points. On the 300 licensed Ocular turns, ElevenLabs Scribe v3 Live scores 1.81% and Soniox STT-RT v5 2.54%, so ElevenLabs Scribe v3 Live makes fewer errors by 0.73 points. Word error rate counts incorrect, missing, and extra words, so lower is better.
ElevenLabs Scribe v3 Live is more accurate on both datasets. On the 1,000 public Pipecat clips, ElevenLabs Scribe v3 Live scores 1.89% and Soniox STT-RT v5 3.02%, so ElevenLabs Scribe v3 Live makes fewer errors by 1.13 points. On the 300 licensed Ocular turns, ElevenLabs Scribe v3 Live scores 1.81% and Soniox STT-RT v5 2.54%, so ElevenLabs Scribe v3 Live makes fewer errors by 0.73 points. Word error rate counts incorrect, missing, and extra words, so lower is better.
Yes. ElevenLabs Scribe v3 Live finishes the transcript 510ms sooner. ElevenLabs Scribe v3 Live has the final transcript 685ms after the speaker stops and Soniox STT-RT v5 1.20s, at the median. Both are measured on the same 206 streamed turns.
ElevenLabs Scribe v3 Live is the stronger pick for voice agents: it makes fewer errors on real conversations and finishes the transcript 510ms 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. ElevenLabs Scribe v3 Live shows first text after 2.05% and Soniox STT-RT v5 after 3.92%, at the median, so ElevenLabs Scribe v3 Live is 1.88 points lower.
Test your voice agent the same way.
Cekura runs these simulated calls on your agent and scores every one.