AssemblyAI STT 3.7 Realtime Agent vs GPT Realtime Whisper

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)
  • STT 3.7 Realtime Agent1.68% · 47ms
  • Realtime Whisper2.16% · 543ms

AssemblyAI STT 3.7 Realtime Agent is ahead of GPT Realtime Whisper on both datasets together, the Pipecat Dataset, time to first text and final-text delay. The Ocular Dataset is too close to call.

Last updated Methodology by Dileep Chagam

Who is better at what

STT 3.7 Realtime AgentRealtime Whisper
  • Overall WER2.41%2.63%
  • Pipecat Dataset WER1.68%2.16%
  • Ocular Dataset WER3.81%3.53%
  • Time to first text493ms1.38s
  • Final-text delay47ms543ms

Lower is better on every measure. Too close to call means the likely ranges overlap.

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.

STT 3.7 Realtime AgentRealtime Whisper
  • ELO rating10241001–1039 · rank 1–1610091005–1020 · rank 5–16
  • Expected win rate52%48%
  • Ocular turns WER · 158 draws of 19828 won12 won
  • Pipecat WER · 900 draws of 99964 won35 won
  • Semantic WER · 188 draws of 1985 won5 won
  • Phone line · 156 draws of 19832 won10 won
  • Call-center noise · 121 draws of 19748 won28 won
  • Babble noise · 112 draws of 19759 won26 won
  • TV noise · 106 draws of 19813 won79 won
  • Reverb · 135 draws of 19855 won8 won
  • Short replies · 53 draws of 7014 won3 won
  • Entities · 47 draws of 502 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

STT 3.7 Realtime Agent vs Realtime Whisper FAQ

  • AssemblyAI STT 3.7 Realtime Agent is more accurate on both datasets. On the 1,000 public Pipecat clips, AssemblyAI STT 3.7 Realtime Agent scores 2.41% and GPT Realtime Whisper 2.63%, so AssemblyAI STT 3.7 Realtime Agent makes fewer errors by 0.21 points. On the 300 licensed Ocular turns, AssemblyAI STT 3.7 Realtime Agent scores 1.68% and GPT Realtime Whisper 2.16%, so AssemblyAI STT 3.7 Realtime Agent makes fewer errors by 0.48 points. Word error rate counts incorrect, missing, and extra words, so lower is better.

  • Yes. AssemblyAI STT 3.7 Realtime Agent finishes the transcript 886ms sooner. AssemblyAI STT 3.7 Realtime Agent has the final transcript 493ms after the speaker stops and GPT Realtime Whisper 1.38s, at the median. Both are measured on the same 206 streamed turns.

  • AssemblyAI STT 3.7 Realtime Agent is the stronger pick for voice agents: it makes fewer errors on real conversations and finishes the transcript 886ms 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. AssemblyAI STT 3.7 Realtime Agent shows first text after 3.81% and GPT Realtime Whisper after 3.53%, at the median.

Test your voice agent the same way.

Cekura runs these simulated calls on your agent and scores every one.