DeepgramCartesiaAudio matchupSpeech to text
Nova-3vs.Cartesia Ink-2
Which model is better?
Nova-3 is slightly ahead in independent tests, but the lead is not certain.
Pick Nova-3 for
Pick Cartesia Ink-2 for
Shared tests
Independent evaluators and leaderboards only, the newest shared snapshot per evaluator. With 95 % ranges a lead counts when they do not overlap; for hit rates, when it holds in a 95 % test. Without any uncertainty figure at most “slightly ahead”.
Coval
Nova-3 slightly aheadNova-3 is clearly ahead in 0 of 16 subtests and slightly in 8, Cartesia Ink-2 in 0 clearly and slightly in 4. 4 are level.
Coval Word Error Rate (WER) per-sample p50
lower is better, as of November 5, 2026
Coval Time to First Token (TTFT) maximum
lower is better, as of November 5, 2026
Coval Time to First Token (TTFT) p50
lower is better, as of November 5, 2026
Show all 16 subtests
| Subtest | Nova-3 | Cartesia Ink-2 | Verdict |
|---|---|---|---|
| Coval Word Error Rate (WER) per-sample p50 (lower is better) | 3.1 % | 0.0 % | Cartesia Ink-2 slightly ahead |
| Coval Time to First Token (TTFT) maximum (lower is better) | 5,457 ms | 25,200 ms | Nova-3 slightly ahead |
| Coval Time to First Token (TTFT) p50 (lower is better) | 1,000 ms | 2,017 ms | Nova-3 slightly ahead |
| Coval Time to Final Segment (TTFS) mean (lower is better) | 86 ms | 145 ms | Nova-3 slightly ahead |
| Coval RTF maximum (lower is better) | 2.89 | 4.71 | Nova-3 slightly ahead |
| Coval Time to Final Segment (TTFS) maximum (lower is better) | 19,227 ms | 31,136 ms | Nova-3 slightly ahead |
| Coval Time to Final Segment (TTFS) p50 (lower is better) | 91 ms | 139 ms | Nova-3 slightly ahead |
| Coval pooled corpus Word Error Rate (WER) (lower is better) | 5.2 % | 3.4 % | Cartesia Ink-2 slightly ahead |
| Coval Word Error Rate (WER) per-sample maximum (lower is better) | 100.0 % | 69.0 % | Cartesia Ink-2 slightly ahead |
| Coval Word Error Rate (WER) per-sample mean (lower is better) | 5.6 % | 3.9 % | Cartesia Ink-2 slightly ahead |
| Coval Time to First Token (TTFT) mean (lower is better) | 1,440 ms | 2,033 ms | Nova-3 slightly ahead |
| Coval AudioToFinal maximum (lower is better) | 29.38 s | 39.52 s | Nova-3 slightly ahead |
| Coval RTF mean (lower is better) | 1.01 | 1.02 | level |
| Coval AudioToFinal mean (lower is better) | 8.02 s | 8.09 s | level |
| Coval AudioToFinal p50 (lower is better) | 8.39 s | 8.46 s | level |
| Coval RTF p50 (lower is better) | 1.01 | 1.02 | level |
Source · latest November 5, 2026
Cekura
Nova-3 slightly aheadNova-3 is clearly ahead in 0 of 2 subtests and slightly in 2, Cartesia Ink-2 in none.
First-text latency (TTFT), P50
lower is better, as of October 23, 2026
Final-text delay (TTFS), P50
lower is better, as of October 23, 2026
Source · latest October 23, 2026
All data side by side
Prices are verbatim from the provider, because audio models bill per minute, character or token.
| Nova-3 | Cartesia Ink-2 | |
|---|---|---|
| Task | Speech to text | Speech to text |
| Developer | Deepgram | Cartesia |
| Status | Active | Active |
| Released | n/a | July 9, 2026 |
| Weights | proprietary | proprietary |
| Runs locally | no | no |
| Price | n/a | 3 credits per second of audio; the dollar price and overage charges depend on the plan |
| Popularity (out of 100) | 51 | 0 |