AssemblyAI Universal 3.5 Pro vs Deepgram Flux English
AssemblyAI Universal 3.5 Pro is ahead on the Pipecat Dataset and time to first text, and Deepgram Flux English on final-text delay. The Ocular Dataset is too close to call.
AssemblyAI Universal 3.5 Pro has a newer release. See AssemblyAI Universal 3.6 Pro vs Deepgram Flux English.
Last updated Methodology by Dileep Chagam
Accuracy vs speed
Each dot is one model. Bottom-left is best: fewer errors and a faster final transcript.
Which to pick
Pick AssemblyAI Universal 3.5 Pro for accuracy on clean, scripted audio (1.97 points lower) and showing words while the caller is still talking (302ms sooner). Pick Deepgram Flux English for quick turn-taking once the caller stops (72ms sooner). For accuracy on real phone conversations, the gap is inside the margin of error, so neither is ahead.
The four metrics, side by side
- AssemblyAI Universal 3.5 Pro
- 1.93%2nd
- Deepgram Flux English
- 3.90%15th
AssemblyAI Universal 3.5 Pro, 1.97 points lower
- AssemblyAI Universal 3.5 Pro
- 4.23%10th
- Deepgram Flux English
- 4.18%8th
Too close to call: inside the margin of error
- AssemblyAI Universal 3.5 Pro
- 489ms2nd
- Deepgram Flux English
- 791ms3rd
AssemblyAI Universal 3.5 Pro, 302ms sooner
- AssemblyAI Universal 3.5 Pro
- 180ms9th
- Deepgram Flux English
- 108ms5th
Deepgram Flux English, 72ms sooner
Lower is better on every metric. The small figure is each model's place in the whole field. What each metric measures.
- WER · Word error rate
- Share of words the model got wrong: replaced, missing or added. Lower is better.
- Pipecat Dataset
- 1,000 short public clips from Pipecat's STT benchmark dataset.
- Ocular Dataset
- 8 recordings from 4 real phone conversations, licensed from Ocular.
- TTFT · Time to first text
- How long after the audio starts streaming until the first words appear.
- TTFS · Final-text delay
- How long after the caller stops until the full transcript arrives. A voice agent waits for this before it replies.
First text and final text, at three percentiles
AssemblyAI Universal 3.5 Pro shows first text after 1.08s and final text after 692ms, and Deepgram Flux English shows first text after 1.51s and final text after 141ms, on a bad turn (P95) across the same streamed turns.
| Measure | Model | P50 | P90 | P95 |
|---|---|---|---|---|
| Time to first text | AssemblyAI Universal 3.5 Pro | 489ms | 966ms | 1.08s |
| Deepgram Flux English | 791ms | 1.23s | 1.51s | |
| Final-text delay | AssemblyAI Universal 3.5 Pro | 180ms | 532ms | 692ms |
| Deepgram Flux English | 108ms | 132ms | 141ms |
- P50 / P90 / P95 · Typical, slow and slowest turns
- Half of turns are faster than P50. Only 1 in 10 is slower than P90, and 1 in 20 slower than P95.
How each ranks among all 17 models
Place among the 17 models on each metric. Lower is better on all four.
| Metric | Universal 3.5 Pro | Flux English |
|---|---|---|
| Pipecat Dataset WER | 1.93%2nd of 17 | 3.90%15th of 17 |
| Ocular Dataset WER | 4.23%10th of 17 | 4.18%8th of 17 |
| Time to first text | 489ms2nd of 15 | 791ms3rd of 15 |
| Final-text delay | 180ms9th of 17 | 108ms5th of 17 |
How much audio each model was tested on
| Model | Pipecat clips | Ocular recordings | Turns timed | Failed |
|---|---|---|---|---|
| AssemblyAI Universal 3.5 Pro | 1,000 | 8 | 205 | 1 |
| Deepgram Flux English | 999 | 8 | 199 | 7 |
Only clips with usable text count, so the two totals can differ. How WER is scored.
Related comparisons
Both providers, every model: AssemblyAI vs Deepgram
- AssemblyAI Universal 3.5 Pro vs AssemblyAI Universal 3.6 Pro
- Deepgram Flux English vs GPT-4o Transcribe
- Deepgram Flux English vs Smallest Pulse
- Deepgram Flux English vs Deepgram Nova-3
- Deepgram Flux English vs GPT-4o Mini Transcribe
- Deepgram Flux English vs Deepgram Flux Multilingual
- Deepgram Flux English vs Inworld STT-1
- Deepgram Flux English vs Soniox STT-RT v5
- Deepgram Flux English vs Speechmatics Linden
- Deepgram Flux English vs Cartesia Ink 2
- Deepgram Flux English vs GPT Realtime Whisper
- Deepgram Flux English vs Reson8
- Deepgram Flux English vs Google Chirp 3
- Deepgram Flux English vs AssemblyAI Universal 3.6 Pro
- Deepgram Flux English vs Gradium
Frequently asked questions
Which is more accurate, AssemblyAI Universal 3.5 Pro or Deepgram Flux English?
AssemblyAI Universal 3.5 Pro is more accurate on the public clips. On the 1,000 public Pipecat clips, AssemblyAI Universal 3.5 Pro scores 1.93% and Deepgram Flux English 3.90%, so AssemblyAI Universal 3.5 Pro makes fewer errors by 1.97 points. On the eight licensed Ocular recordings, AssemblyAI Universal 3.5 Pro scores 4.23% and Deepgram Flux English 4.18%, with likely ranges of 3.57 to 4.68% and 3.08 to 5.23%. The ranges overlap, so the difference is too close to call. Word error rate counts incorrect, missing, and extra words, so lower is better.
Is AssemblyAI Universal 3.5 Pro faster than Deepgram Flux English?
No. Deepgram Flux English finishes the transcript 72ms sooner. AssemblyAI Universal 3.5 Pro has the final transcript 180ms after the speaker stops and Deepgram Flux English 108ms, at the median. Both are measured on the same 206 streamed turns.
Is AssemblyAI Universal 3.5 Pro or Deepgram Flux English better for voice agents?
Deepgram Flux English finishes the transcript 72ms 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.
Do AssemblyAI Universal 3.5 Pro and Deepgram Flux English stream text while the speaker is talking?
Yes, both return partial text before the speaker stops. AssemblyAI Universal 3.5 Pro shows first text after 489ms and Deepgram Flux English after 791ms, at the median, so AssemblyAI Universal 3.5 Pro is 302ms sooner.