Deepgram Nova-3 vs GPT Realtime Whisper
Deepgram Nova-3 is ahead on time to first text and final-text delay, and GPT Realtime Whisper on the Pipecat Dataset. The Ocular Dataset is too close to call.
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 Deepgram Nova-3 for showing words while the caller is still talking (315ms sooner) and quick turn-taking once the caller stops (440ms sooner). Pick GPT Realtime Whisper for accuracy on clean, scripted audio (1.31 points lower). For accuracy on real phone conversations, the gap is inside the margin of error, so neither is ahead.
The four metrics, side by side
- Deepgram Nova-3
- 3.46%12th
- GPT Realtime Whisper
- 2.16%5th
GPT Realtime Whisper, 1.31 points lower
- Deepgram Nova-3
- 4.39%11th
- GPT Realtime Whisper
- 3.53%4th
Too close to call: inside the margin of error
- Deepgram Nova-3
- 1.06s7th
- GPT Realtime Whisper
- 1.38s10th
Deepgram Nova-3, 315ms sooner
- Deepgram Nova-3
- 103ms4th
- GPT Realtime Whisper
- 543ms14th
Deepgram Nova-3, 440ms 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
Deepgram Nova-3 shows first text after 2.07s and final text after 155ms, and GPT Realtime Whisper shows first text after 2.86s and final text after 669ms, on a bad turn (P95) across the same streamed turns.
| Measure | Model | P50 | P90 | P95 |
|---|---|---|---|---|
| Time to first text | Deepgram Nova-3 | 1.06s | 1.10s | 2.07s |
| GPT Realtime Whisper | 1.38s | 1.84s | 2.86s | |
| Final-text delay | Deepgram Nova-3 | 103ms | 143ms | 155ms |
| GPT Realtime Whisper | 543ms | 647ms | 669ms |
- 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 | Nova-3 | Realtime Whisper |
|---|---|---|
| Pipecat Dataset WER | 3.46%12th of 17 | 2.16%5th of 17 |
| Ocular Dataset WER | 4.39%11th of 17 | 3.53%4th of 17 |
| Time to first text | 1.06s7th of 15 | 1.38s10th of 15 |
| Final-text delay | 103ms4th of 17 | 543ms14th of 17 |
How much audio each model was tested on
| Model | Pipecat clips | Ocular recordings | Turns timed | Failed |
|---|---|---|---|---|
| Deepgram Nova-3 | 1,000 | 8 | 202 | 0 |
| GPT Realtime Whisper | 1,000 | 8 | 206 | 0 |
Only clips with usable text count, so the two totals can differ. How WER is scored.
Related comparisons
Both providers, every model: Deepgram vs OpenAI
- Deepgram Nova-3 vs GPT-4o Mini Transcribe
- Deepgram Nova-3 vs Smallest Pulse
- Deepgram Nova-3 vs GPT-4o Transcribe
- Deepgram Nova-3 vs Deepgram Flux English
- Deepgram Nova-3 vs Inworld STT-1
- Deepgram Nova-3 vs Soniox STT-RT v5
- Deepgram Nova-3 vs Speechmatics Linden
- Deepgram Nova-3 vs Cartesia Ink 2
- Deepgram Nova-3 vs Reson8
- Deepgram Nova-3 vs Google Chirp 3
- Deepgram Nova-3 vs Deepgram Flux Multilingual
- Deepgram Nova-3 vs AssemblyAI Universal 3.6 Pro
- Deepgram Nova-3 vs Gradium
- GPT Realtime Whisper vs Reson8
- GPT Realtime Whisper vs Google Chirp 3
- GPT Realtime Whisper vs Cartesia Ink 2
- GPT Realtime Whisper vs Speechmatics Linden
- GPT Realtime Whisper vs AssemblyAI Universal 3.6 Pro
- GPT Realtime Whisper vs Soniox STT-RT v5
- GPT Realtime Whisper vs Inworld STT-1
- GPT Realtime Whisper vs GPT-4o Mini Transcribe
- GPT Realtime Whisper vs Smallest Pulse
- GPT Realtime Whisper vs GPT-4o Transcribe
- GPT Realtime Whisper vs Deepgram Flux English
- GPT Realtime Whisper vs Deepgram Flux Multilingual
- GPT Realtime Whisper vs Gradium
Frequently asked questions
Which is more accurate, Deepgram Nova-3 or GPT Realtime Whisper?
GPT Realtime Whisper is more accurate on the public clips. On the 1,000 public Pipecat clips, Deepgram Nova-3 scores 3.46% and GPT Realtime Whisper 2.16%, so GPT Realtime Whisper makes fewer errors by 1.31 points. On the eight licensed Ocular recordings, Deepgram Nova-3 scores 4.39% and GPT Realtime Whisper 3.53%, with likely ranges of 3.42 to 5.25% and 2.58 to 4.51%. 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 Deepgram Nova-3 faster than GPT Realtime Whisper?
Yes. Deepgram Nova-3 finishes the transcript 440ms sooner. Deepgram Nova-3 has the final transcript 103ms after the speaker stops and GPT Realtime Whisper 543ms, at the median. Both are measured on the same 206 streamed turns.
Is Deepgram Nova-3 or GPT Realtime Whisper better for voice agents?
Deepgram Nova-3 finishes the transcript 440ms 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 Deepgram Nova-3 and GPT Realtime Whisper stream text while the speaker is talking?
Yes, both return partial text before the speaker stops. Deepgram Nova-3 shows first text after 1.06s and GPT Realtime Whisper after 1.38s, at the median, so Deepgram Nova-3 is 315ms sooner.