GPT-4o Transcribe vs Soniox STT-RT v5

Compare

Accuracy and speed

Left gets fewer words wrong. Lower has the final text ready sooner.

Accurate and fast0ms200ms400ms600ms800ms0%2%4%6%Words wrong on clean audio (Pipecat WER)
  • 4o Transcribe3.90% · 652ms
  • STT-RT v52.54% · 38ms

Soniox STT-RT v5 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

4o TranscribeSTT-RT v5
  • Overall WER6.85%3.02%
  • Pipecat Dataset WER3.90%2.54%
  • Ocular Dataset WER12.45%3.92%
  • Time to first textn/a1.20s
  • Final-text delay652ms38ms

Lower is better on every measure. Too close to call means the likely ranges overlap. Our guide covers testing OpenAI Realtime API voice agents.

Clip by clip

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.

4o TranscribeSTT-RT v5
  • ELO rating873834–934 · rank 22–231000986–1007 · rank 6–20
  • Expected win rate33%67%
  • Ocular turns WER · 100 draws of 19419 won75 won
  • Pipecat WER · 785 draws of 99984 won130 won
  • Semantic WER · 154 draws of 1947 won33 won
  • Phone line · 95 draws of 19817 won86 won
  • Call-center noise · 64 draws of 19822 won112 won
  • Babble noise · 60 draws of 19830 won108 won
  • TV noise · 86 draws of 19865 won47 won
  • Reverb · 77 draws of 19823 won98 won
  • Short replies · 30 draws of 6911 won28 won
  • Entities · 45 draws of 482 won1 won

Short replies are decided on the error count and entities on exact capture, with no draw margin. How the rating is worked out

4o Transcribe vs STT-RT v5 FAQ

  • Soniox STT-RT v5 is more accurate on both datasets. On the 1,000 public Pipecat clips, GPT-4o Transcribe scores 6.85% and Soniox STT-RT v5 3.02%, so Soniox STT-RT v5 makes fewer errors by 3.83 points. On the 300 licensed Ocular turns, GPT-4o Transcribe scores 3.90% and Soniox STT-RT v5 2.54%, so Soniox STT-RT v5 makes fewer errors by 1.36 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.

  • Soniox STT-RT v5 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 Soniox STT-RT v5. 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.