GPT-4o Mini TranscribevsZoom Scribe
Accuracy and speed
Left gets fewer words wrong. Lower has the final text ready sooner.
- 4o Mini Transcribe3.35% · 630ms
- Scribe1.97% · 118ms
Zoom Scribe is ahead of GPT-4o Mini Transcribe on the Pipecat Dataset and final-text delay. The Ocular Dataset is too close to call.
Last updated Methodology by Dileep Chagam
Who is better at what
- Pipecat Dataset WER3.35%1.97%
- Ocular Dataset WER4.45%4.88%
- Time to first textn/a1.67s
- Final-text delay630ms118ms
Lower is better on every measure. Too close to call means the likely ranges overlap.
Related comparisons
Both providers, every model: OpenAI vs Zoom
4o Mini Transcribe vs Scribe FAQ
Which is more accurate, GPT-4o Mini Transcribe or Zoom Scribe?
Zoom Scribe is more accurate on the public clips. On the 1,000 public Pipecat clips, GPT-4o Mini Transcribe scores 3.35% and Zoom Scribe 1.97%, so Zoom Scribe makes fewer errors by 1.39 points. On the eight licensed Ocular recordings, GPT-4o Mini Transcribe scores 4.45% and Zoom Scribe 4.88%, with likely ranges of 3.22 to 5.55% and 4.41 to 5.29%. The ranges overlap, so the difference is too close to call. Word error rate counts incorrect, missing, and extra words, so lower is better.
Zoom Scribe is more accurate on the public clips. On the 1,000 public Pipecat clips, GPT-4o Mini Transcribe scores 3.35% and Zoom Scribe 1.97%, so Zoom Scribe makes fewer errors by 1.39 points. On the eight licensed Ocular recordings, GPT-4o Mini Transcribe scores 4.45% and Zoom Scribe 4.88%, with likely ranges of 3.22 to 5.55% and 4.41 to 5.29%. The ranges overlap, so the difference is too close to call. Word error rate counts incorrect, missing, and extra words, so lower is better.
No. Zoom Scribe finishes the transcript 512ms sooner. GPT-4o Mini Transcribe has the final transcript 630ms after the speaker stops and Zoom Scribe 118ms, at the median. Both are measured on the same 206 streamed turns.
Zoom Scribe finishes the transcript 512ms sooner, and the two are too close to call on real conversations. A voice agent cannot reply until the final transcript arrives, and the Ocular recordings are real conversations.
Only Zoom Scribe. GPT-4o Mini Transcribe returns no text until the speaker stops, which matters if your agent reacts to partial transcripts.