Cartesia Ink 2 vs Deepgram Nova-3
Cartesia Ink 2 is ahead on the Pipecat Dataset and the Ocular Dataset, and Deepgram Nova-3 on time to first text and final-text delay.
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.
- Cartesia Ink 2
- 2.39%5th
- Deepgram Nova-3
- 3.46%10th
Cartesia Ink 2, 1.08 points lower
Word error rate on eight licensed recordings provided by Ocular.
- Cartesia Ink 2
- 3.09%2nd
- Deepgram Nova-3
- 4.39%10th
Cartesia Ink 2, 1.31 points lower
Median delay before the first words arrive, from the first audio packet.
- Cartesia Ink 2
- 1.51s9th
- Deepgram Nova-3
- 1.06s6th
Deepgram Nova-3, 447ms sooner
Median delay after the speaker stops before the transcript is final.
- Cartesia Ink 2
- 123ms5th
- Deepgram Nova-3
- 103ms2nd
Deepgram Nova-3, 20ms 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 | Cartesia Ink 2 | 1.51s | 2.04s | 3.23s |
| Deepgram Nova-3 | 1.06s | 1.10s | 2.07s | |
| Final-text delay | Cartesia Ink 2 | 123ms | 139ms | 145ms |
| Deepgram Nova-3 | 103ms | 143ms | 155ms |
Both against every other model
Open on the metric where the two sit furthest apart, with the other three a tab away.
Word error rate on eight licensed recordings provided by Ocular. Lower is better.
| # | Model | Ocular Dataset | clips | |
|---|---|---|---|---|
| 1 | Reson8 | 2.93% | 8 | |
| 2 | Cartesia Ink 2 | 3.09% | 8 | |
| 3 | Smallest Pulse | 3.46% | 8 | |
| 4 | GPT Realtime Whisper | 3.53% | 8 | |
| 5 | Google Chirp 2 | 3.80% | 8 | |
| 6 | Speechmatics Linden | 3.93% | 8 | |
| 7 | Deepgram Flux English | 4.18% | 8 | |
| 8 | Google Chirp 3 | 4.19% | 8 | |
| 9 | AssemblyAI Universal 3.5 Pro | 4.23% | 8 | |
| 10 | Deepgram Nova-3 | 4.39% | 8 | |
| 11 | Gradium | 4.45% | 8 | |
| 12 | GPT-4o Mini Transcribe | 4.45% | 8 | |
| 13 | Deepgram Flux Multilingual | 4.53% | 8 | |
| 14 | Inworld STT-1 | 10.60% | 8 | |
| 15 | GPT-4o Transcribe | 12.45% | 8 |
Counts are the clips each model returned usable text for, so they differ between models and between datasets.
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 |
|---|---|---|---|---|
| Cartesia Ink 2 | 1,000 | 8 | 198 | 0 |
| Deepgram Nova-3 | 1,000 | 8 | 202 | 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 Cartesia Ink 2 or Deepgram Nova-3 more accurate?
On the 1,000 public Pipecat clips, Cartesia Ink 2 scores 2.39% and Deepgram Nova-3 3.46%, so Cartesia Ink 2 makes fewer errors by 1.08 points. On the eight licensed Ocular recordings, Cartesia Ink 2 scores 3.09% and Deepgram Nova-3 4.39%, so Cartesia Ink 2 makes fewer errors by 1.31 points. Word error rate counts incorrect, missing, and extra words against the reference transcript, so lower is better.
Which is faster, Cartesia Ink 2 or Deepgram Nova-3?
Deepgram Nova-3 returns first text 447ms sooner, at 1.06s against 1.51s at the median. Deepgram Nova-3 has the final transcript 20ms sooner, 103ms after the speaker stops against 123ms. Both are measured on the same 206 streamed turns.
How were Cartesia Ink 2 and Deepgram Nova-3 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. Cartesia Ink 2 returned usable text for 1,000 of the public clips and Deepgram Nova-3 for 1,000. Latency uses each provider's own supported finalization contract.