AssemblyAI STT 3.7 Realtime Agent vs AssemblyAI Universal 3.6 Pro

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
  • Universal 3.6 Pro1.77% · 91ms

AssemblyAI STT 3.7 Realtime Agent is ahead on both datasets together and final-text delay, and AssemblyAI Universal 3.6 Pro on time to first text. The Pipecat Dataset and The Ocular Dataset are too close to call.

Last updated Methodology by Dileep Chagam

Who is better at what

STT 3.7 Realtime AgentUniversal 3.6 Pro
  • Overall WER2.41%2.71%
  • Pipecat Dataset WER1.68%1.77%
  • Ocular Dataset WER3.81%4.50%
  • Time to first text493ms485ms
  • Final-text delay47ms91ms

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 AgentUniversal 3.6 Pro
  • ELO rating10241001–1039 · rank 1–1610211003–1031 · rank 2–16
  • Expected win rate50%50%
  • Ocular turns WER · 180 draws of 19612 won4 won
  • Pipecat WER · 968 draws of 99922 won9 won
  • Semantic WER · 194 draws of 1961 won1 won
  • Phone line · 190 draws of 1986 won2 won
  • Call-center noise · 159 draws of 19718 won20 won
  • Babble noise · 167 draws of 19720 won10 won
  • TV noise · 174 draws of 19819 won5 won
  • Reverb · 181 draws of 19810 won7 won
  • Short replies · 64 draws of 683 won1 won
  • Entities · 49 draws of 501 won0 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 Universal 3.6 Pro FAQ

  • AssemblyAI STT 3.7 Realtime Agent is more accurate on the public clips. On the 1,000 public Pipecat clips, AssemblyAI STT 3.7 Realtime Agent scores 2.41% and AssemblyAI Universal 3.6 Pro 2.71%, so AssemblyAI STT 3.7 Realtime Agent makes fewer errors by 0.30 points. On the 300 licensed Ocular turns, AssemblyAI STT 3.7 Realtime Agent scores 1.68% and AssemblyAI Universal 3.6 Pro 1.77%, with likely ranges of 1.47 to 1.90% and 1.55 to 2.00%. 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. AssemblyAI Universal 3.6 Pro finishes the transcript 8ms sooner. AssemblyAI STT 3.7 Realtime Agent has the final transcript 493ms after the speaker stops and AssemblyAI Universal 3.6 Pro 485ms, at the median. Both are measured on the same 206 streamed turns.

  • AssemblyAI Universal 3.6 Pro finishes the transcript 8ms 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.

  • Yes, both return partial text before the speaker stops. AssemblyAI STT 3.7 Realtime Agent shows first text after 3.81% and AssemblyAI Universal 3.6 Pro after 4.50%, at the median.

Test your voice agent the same way.

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