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 and the Ocular Dataset.
Quality and speed frontier
Every model's error rate on the public clips, against how long its transcript takes to settle.
The four metrics, side by side
Word error rate on 1,000 clips from Pipecat's STT benchmark dataset.
- Deepgram Nova-3
- 3.46%10th
- GPT Realtime Whisper
- 2.16%4th
GPT Realtime Whisper, 1.31 points lower
Word error rate on eight licensed recordings provided by Ocular.
- Deepgram Nova-3
- 4.39%10th
- GPT Realtime Whisper
- 3.53%4th
GPT Realtime Whisper, 0.87 points lower
Median delay before the first words arrive, from the first audio packet.
- Deepgram Nova-3
- 1.06s6th
- GPT Realtime Whisper
- 1.38s8th
Deepgram Nova-3, 315ms sooner
Median delay after the speaker stops before the transcript is final.
- Deepgram Nova-3
- 103ms2nd
- GPT Realtime Whisper
- 543ms12th
Deepgram Nova-3, 440ms sooner
Lower is better on every metric. The small figure is each model's place in the whole field.
First text and final text, at three percentiles
Both measured on the same 206 streamed Ocular Dataset turns. P90 and P95 say what a caller meets on a bad turn, which a median hides.
| 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 |
Both against every other model
Open on the metric where the two sit furthest apart, with the other three a tab away.
Final-text delay after the speaker stops. Lower is better.
| # | Model | Final text | turns | |
|---|---|---|---|---|
| 1 | Inworld STT-1 | 41ms | 206 | |
| 2 | Deepgram Nova-3 | 103ms | 202 | |
| 3 | Deepgram Flux English | 108ms | 199 | |
| 4 | Deepgram Flux Multilingual | 114ms | 175 | |
| 5 | Cartesia Ink 2 | 123ms | 198 | |
| 6 | Speechmatics Linden | 147ms | 194 | |
| 7 | AssemblyAI Universal 3.5 Pro | 180ms | 205 | |
| 8 | Smallest Pulse | 201ms | 206 | |
| 9 | Gradium | 212ms | 198 | |
| 10 | Reson8 | 235ms | 201 | |
| 11 | Google Chirp 3 | 475ms | 206 | |
| 12 | GPT Realtime Whisper | 543ms | 206 | |
| 13 | Google Chirp 2 | 568ms | 205 | |
| 14 | GPT-4o Mini Transcribe | 630ms | 206 | |
| 15 | GPT-4o Transcribe | 652ms | 206 |
Median across the same 206 Ocular Dataset turns. Providers finalize differently, so compare models with similar finalization contracts before treating small gaps as meaningful.
Both on the other two views
Pipecat vs Ocular accuracy
Every model's error rate on the public clips, against the same model on real conversations.
Transcription response time
How soon each model returns its first words, against how soon its transcript is final.
What the numbers rest 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 |
A clip counts only where the model returned usable text, so counts differ between models. Word error rate is measured the same way throughout: incorrect, missing, and extra words against the reference transcript.
Frequently asked questions
Is Deepgram Nova-3 or GPT Realtime Whisper more accurate?
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%, so GPT Realtime Whisper makes fewer errors by 0.87 points. Word error rate counts incorrect, missing, and extra words against the reference transcript, so lower is better.
Which is faster, Deepgram Nova-3 or GPT Realtime Whisper?
Deepgram Nova-3 returns first text 315ms sooner, at 1.06s against 1.38s at the median. Deepgram Nova-3 has the final transcript 440ms sooner, 103ms after the speaker stops against 543ms. Both are measured on the same 206 streamed turns.
How were Deepgram Nova-3 and GPT Realtime Whisper compared?
Both ran the same two datasets under the same conditions as every other model in the benchmark: 1,000 public clips from Pipecat's STT benchmark dataset and eight licensed conversation recordings from Ocular. Deepgram Nova-3 returned usable text for 1,000 of the public clips and GPT Realtime Whisper for 1,000. Latency uses each provider's own supported finalization contract.