Fish Audio Transcribe Stream 1 vs GPT-4o Transcribe
Accuracy and speed
Left gets fewer words wrong. Lower has the final text ready sooner.
- Transcribe Stream 11.90% · 123ms
- 4o Transcribe3.90% · 652ms
Fish Audio Transcribe Stream 1 is ahead of GPT-4o Transcribe on both datasets together, the Pipecat Dataset, the Ocular Dataset and final-text delay.
Last updated Methodology by Dileep Chagam
Who is better at what
- Overall WER2.21%6.85%
- Pipecat Dataset WER1.90%3.90%
- Ocular Dataset WER2.78%12.45%
- Time to first text730msn/a
- Final-text delay123ms652ms
Lower is better on every measure. Too close to call means the likely ranges overlap. Our guide covers testing OpenAI Realtime API voice agents.
Transcribe Stream 1 vs 4o Transcribe FAQ
Which is more accurate, Fish Audio Transcribe Stream 1 or GPT-4o Transcribe?
Fish Audio Transcribe Stream 1 is more accurate on both datasets. On the 1,000 public Pipecat clips, Fish Audio Transcribe Stream 1 scores 2.21% and GPT-4o Transcribe 6.85%, so Fish Audio Transcribe Stream 1 makes fewer errors by 4.65 points. On the eight licensed Ocular recordings, Fish Audio Transcribe Stream 1 scores 1.90% and GPT-4o Transcribe 3.90%, so Fish Audio Transcribe Stream 1 makes fewer errors by 2.00 points. Word error rate counts incorrect, missing, and extra words, so lower is better.
Fish Audio Transcribe Stream 1 is more accurate on both datasets. On the 1,000 public Pipecat clips, Fish Audio Transcribe Stream 1 scores 2.21% and GPT-4o Transcribe 6.85%, so Fish Audio Transcribe Stream 1 makes fewer errors by 4.65 points. On the eight licensed Ocular recordings, Fish Audio Transcribe Stream 1 scores 1.90% and GPT-4o Transcribe 3.90%, so Fish Audio Transcribe Stream 1 makes fewer errors by 2.00 points. 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.
Fish Audio Transcribe Stream 1 makes fewer errors on real conversations, and the two are not separated on speed. A voice agent cannot reply until the final transcript arrives, and the Ocular recordings are real conversations.
Only Fish Audio Transcribe Stream 1. 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.