Left gets fewer words wrong. Lower has the final text ready sooner.
ElevenLabs Scribe v3 Live is ahead on both datasets together, the Pipecat Dataset, the Ocular Dataset and time to first text, and Murf Finch-Global on final-text delay.
Last updated Methodology by Dileep Chagam
Lower is better on every measure. Too close to call means the likely ranges overlap.
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.
Short replies are decided on the error count and entities on exact capture, with no draw margin. How the rating is worked out
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 Murf Finch-Global 3.33%, so ElevenLabs Scribe v3 Live makes fewer errors by 1.43 points. On the 300 licensed Ocular turns, ElevenLabs Scribe v3 Live scores 1.81% and Murf Finch-Global 3.17%, so ElevenLabs Scribe v3 Live makes fewer errors by 1.36 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 Murf Finch-Global 3.33%, so ElevenLabs Scribe v3 Live makes fewer errors by 1.43 points. On the 300 licensed Ocular turns, ElevenLabs Scribe v3 Live scores 1.81% and Murf Finch-Global 3.17%, so ElevenLabs Scribe v3 Live makes fewer errors by 1.36 points. Word error rate counts incorrect, missing, and extra words, so lower is better.
Yes. ElevenLabs Scribe v3 Live finishes the transcript 1.15s sooner. ElevenLabs Scribe v3 Live has the final transcript 685ms after the speaker stops and Murf Finch-Global 1.83s, 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 1.15s 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 Murf Finch-Global after 3.62%, at the median, so ElevenLabs Scribe v3 Live is 1.58 points lower.
Cekura runs these simulated calls on your agent and scores every one.