GPT-4o Transcribe vs Smallest Pulse
Accuracy and speed
Left gets fewer words wrong. Lower has the final text ready sooner.
- 4o Transcribe3.90% · 652ms
- Pulse3.66% · 201ms
Smallest Pulse is ahead of GPT-4o Transcribe on both datasets together, the Ocular Dataset and final-text delay. The Pipecat Dataset is too close to call.
Last updated Methodology by Dileep Chagam
Who is better at what
- Overall WER6.85%3.59%
- Pipecat Dataset WER3.90%3.66%
- Ocular Dataset WER12.45%3.46%
- Time to first textn/a1.06s
- Final-text delay652ms201ms
Lower is better on every measure. Too close to call means the likely ranges overlap. Our guide covers testing OpenAI Realtime API voice agents.
4o Transcribe vs Pulse FAQ
Which is more accurate, GPT-4o Transcribe or Smallest Pulse?
Smallest Pulse is more accurate on the public clips. On the 1,000 public Pipecat clips, GPT-4o Transcribe scores 6.85% and Smallest Pulse 3.59%, so Smallest Pulse makes fewer errors by 3.27 points. On the eight licensed Ocular recordings, GPT-4o Transcribe scores 3.90% and Smallest Pulse 3.66%, with likely ranges of 3.43 to 4.42% and 3.34 to 3.99%. The ranges overlap, so the difference is too close to call. Word error rate counts incorrect, missing, and extra words, so lower is better.
Smallest Pulse is more accurate on the public clips. On the 1,000 public Pipecat clips, GPT-4o Transcribe scores 6.85% and Smallest Pulse 3.59%, so Smallest Pulse makes fewer errors by 3.27 points. On the eight licensed Ocular recordings, GPT-4o Transcribe scores 3.90% and Smallest Pulse 3.66%, with likely ranges of 3.43 to 4.42% and 3.34 to 3.99%. The ranges overlap, so the difference is too close to call. Word error rate counts incorrect, missing, and extra words, so lower is better.
Final-text delay was not measured for one or both, so speed cannot be compared. Both are measured on the same 206 streamed turns.
GPT-4o Transcribe and Smallest Pulse are level on the two figures a voice agent depends on most. A voice agent cannot reply until the final transcript arrives, and the Ocular recordings are real conversations.
Only Smallest Pulse. GPT-4o Transcribe returns no text until the speaker stops, which matters if your agent reacts to partial transcripts.
Test your voice agent the same way.
Cekura runs these simulated calls on your agent and scores every one.