Deepgram · read on October 27, 2026
Deepgram Aura 2
Deepgram Aura 2 is a TTS (text-to-speech) model from Deepgram. Coval measures its P95 time to first audio at 525.4 ms, which puts it in the Visible delay speed tier. It ranks 24 of 86 in the full ranking.
- Speed tier
- Visible delay
- P95 / median
- 525.4 ms / 263 ms
- Listener Elo
- Not rated
- Price per 1M chars
- $30
Where it ranks
Full ranking: 24 of 86. Artificial Analysis does not rate this model for a use case, and it is not among the fastest, so it appears on no use case page.
The tier and the ranks are computed by this site from the benchmarks' figures, not declared by Deepgram.
Measurements and sources
Coval
| Median time to first audio (P50) | 263 ms |
| Time to first audio, P95 | 525.4 ms |
| Time to first audio, mean | 299.1 ms |
| Time to first audio, P90 | 492.2 ms |
| Timed requests | 2,808 |
Speko
| Naturalness score | 1,584 |
| Latency, board chart | 147 ms |
| Latency, model page sweep (median) | 125 ms |
| Numbers and dates pronunciation | 0.2 |
Cekura
| Time to first audio, streamed text | 126 ms |
| Time to first audio, text sent whole | 281 ms |
| Reading accuracy | 77.2% |
| Estimated naturalness (1 to 5) | 4.43 |
Details
| Provider | Deepgram |
| Access | Proprietary |
| Streaming | A benchmark measures its time to first audio |
| Languages tested | Spanish, German, French |
| Measured by | Cekura, Coval, Speko |
| Official site | deepgram.com |
Not collected
Nothing below is estimated. Each line says which measure is missing and why.
- Listener rating (Artificial Analysis Speech Arena Elo): not measured by Artificial Analysis (reading of 2026-10-27)
- Price (Artificial Analysis price per 1M characters): not measured by Artificial Analysis (reading of 2026-10-27)
Sources
- https://benchmarks.cekura.ai/text-to-speech (read on October 27, 2026)
- https://benchmarks-api-6wxgp27p2a-ue.a.run.app/v1/results/aggregates?benchmark=TTS&window=30d&include_series=false (read on October 27, 2026)
- https://benchmarks.speko.ai/tts/ (read on October 27, 2026)
Values are copied from these pages on the date shown. Vendor and model name agree in every source that lists it.
Compare it with the reference model: Gradium TTS vs Deepgram Aura 2.