Voice AI BenchmarksBy

PROVIDERLAST 30 DAYS

official resource

Deepgram voice AI models and benchmarks

Deepgram lists 7 STT and TTS models in Coval. Fastest dated mean latency over 30 days: STT: Nova 3 at 86 ms TTFS, with 5.3% WER. TTS: Flux TTS at 205 ms TTFA, with 1.7% WER. Coval benchmarks hosted endpoints daily using a consistent evaluation methodology. Last measured .

Deepgram develops and hosts speech recognition and synthesis APIs.

Measured models
7
STTTTS
Best STT model#5 / 29
86ms
Nova 3
Best TTS model#13 / 32
205ms
Flux TTS

Overview

Deepgram's lineup spans the Nova and Flux recognition families and Aura synthesis, with English, multilingual and voice variants offered as distinct models.

Every model below is measured daily on the same fixed inputs and ranked against the full field, never blended into a company score.

Model lineup

Speech-to-Text
Nova 2, Nova 3, Flux, Flux Multilingual
Text-to-Speech
Aura 2 En, Aura 2, Flux TTS

Speech-to-Text

Full STT dashboard

Ranked on Time to Final Segment against 29 measured models.

Time to Final Segmentms · lower is better · Deepgram models markedEvery measured STT model on Time to Final Segment, with Deepgram's models highlighted.
  1. #1Qwen3 ASR 1.7b39 ms
  2. #2STT RT v545 ms
  3. #3Qwen3 ASR Fast47 ms
  4. #4STT 167 ms
  5. #5Nova 386 ms
  6. #6Nova 287 ms
  7. #7Parakeet TDT 0.6B v397 ms
  8. #8Flux Multilingual98 ms
  9. #9Flux99 ms
leaders plus Deepgram models · 20 other models in the full table
Word Error Rate% · lower is better · Deepgram models markedEvery measured STT model on Word Error Rate, with Deepgram's models highlighted.
  1. #1Universal 3.6 Pro2.1%
  2. #2Universal 3.5 Pro2.4%
  3. #3Qwen3 ASR Fast2.6%
  4. #4resonant-12.6%
  5. #5Qwen3 ASR 1.7b2.9%
  6. #6Gemini 3.5 Transcribe Live3.0%
  7. #7Chirp 33.2%
  8. #20Nova 35.3%
  9. #22Flux5.8%
  10. #25Nova 26.7%
  11. #28Flux Multilingual7.5%
leaders plus Deepgram models · 20 other models in the full table
Time to First Tokenms · lower is better · Deepgram models markedEvery measured STT model on Time to First Token, with Deepgram's models highlighted.
  1. #1Whisper Large v3via Baseten924 ms
  2. #2Qwen3 ASR 1.7b946 ms
  3. #3Universal 3.5 Pro1026 ms
  4. #4Universal 3.6 Pro1071 ms
  5. #5Flux1091 ms
  6. #6Flux Multilingual1165 ms
  7. #7STT 11205 ms
  8. #10Nova 21417 ms
  9. #11Nova 31432 ms
leaders plus Deepgram models · 18 other models in the full table
TTFS vs WER — Deepgram vs every measured modelEach point is one measured STT model · Deepgram models highlighted · 30-day averagesTime to Final Segment against Word Error Rate for every measured STT model, with Deepgram's models highlighted.
Qwen3 ASR 1.7b — 39 ms, 2.9%STT RT v5 — 45 ms, 4.8%Qwen3 ASR Fast — 47 ms, 2.6%STT 1 — 67 ms, 7.1%Parakeet TDT 0.6B v3 — 97 ms, 11.1%Scribe v2 Realtime — 122 ms, 4.4%Whisper Large v3 via Baseten — 131 ms, 4.5%Universal 3.6 Pro — 135 ms, 2.1%Ink 2 — 149 ms, 3.6%Universal 3.5 Pro — 170 ms, 2.4%Whisper Large v3 via Together AI — 198 ms, 7.4%Pulse — 205 ms, 4.6%Grok STT — 223 ms, 4.1%resonant-1 — 252 ms, 2.6%Default — 261 ms, 8.8%Linden 1 — 274 ms, 3.9%Gemini 3.5 Transcribe Live — 317 ms, 3.0%Daytona STT — 399 ms, 4.9%Voxtral Mini Transcribe Realtime 2602 — 415 ms, 5.4%GPT Realtime Whisper — 599 ms, 4.0%Solaria 1 — 620 ms, 6.1%GPT-4o mini Transcribe — 710 ms, 3.9%Chirp 3 — 768 ms, 3.2%GPT-4o Transcribe — 784 ms, 3.8%Chirp 2 — 909 ms, 3.9%Nova 3 — 86 ms, 5.3%Nova 2 — 87 ms, 6.7%Flux Multilingual — 98 ms, 7.5%Flux — 99 ms, 5.8%
Benchmarked models with their Time to Final Segment over the last 30 days.
ModelHostTTFSRank
Nova 2Deepgram87 ms6th of 29
Nova 3Deepgram86 ms5th of 29
FluxDeepgram99 ms9th of 29
Flux MultilingualDeepgram98 ms8th of 29

Text-to-Speech

Full TTS dashboard

Ranked on Time to First Audio against 32 measured models.

Time to First Audioms · lower is better · Deepgram models markedEvery measured TTS model on Time to First Audio, with Deepgram's models highlighted.
  1. #1vui59 ms
  2. #2Default63 ms
  3. #3Qwen3 TTS Fast68 ms
  4. #4TTS Flash 291 ms
  5. #5Qwen3 TTS 1.7b107 ms
  6. #6Simba 3.0111 ms
  7. #7Palabra TTS120 ms
  8. #13Flux TTS205 ms
  9. #19Aura 2297 ms
  10. #28Aura 2 En569 ms
leaders plus Deepgram models · 22 other models in the full table
Word Error Rate% · lower is better · Deepgram models markedEvery measured TTS model on Word Error Rate, with Deepgram's models highlighted.
  1. #1Sonic 3.61.5%
  2. #2S2.1 Pro Free1.5%
  3. #3Simba 3.01.5%
  4. #4S2.1 Pro1.5%
  5. #5Lightning v3.1 Pro1.5%
  6. #6Sonic 3.51.5%
  7. #7Chirp 3 HD1.5%
  8. #13Flux TTS1.7%
  9. #21Aura 21.8%
  10. #25Aura 2 En2.1%
leaders plus Deepgram models · 22 other models in the full table
TTFA vs WER — Deepgram vs every measured modelEach point is one measured TTS model · Deepgram models highlighted · 30-day averagesTime to First Audio against Word Error Rate for every measured TTS model, with Deepgram's models highlighted.
vui — 59 ms, 2.1%Default — 63 ms, 1.9%Qwen3 TTS Fast — 68 ms, 1.8%TTS Flash 2 — 91 ms, 1.7%Qwen3 TTS 1.7b — 107 ms, 1.8%Simba 3.0 — 111 ms, 1.5%Palabra TTS — 120 ms, 2.6%Simba 3.2 — 126 ms, 1.9%drama3-preview — 167 ms, 2.0%TTS 2 — 191 ms, 1.6%Flash v2.5 — 192 ms, 1.8%Eleven v4 Turbo — 201 ms, 1.7%TTS RT v1 — 255 ms, 2.4%TTS Rt v2 — 256 ms, 2.4%Mist v3 — 287 ms, 2.1%Phantom Z 3.4 conversational — 291 ms, 1.8%Sonic 3.5 — 293 ms, 1.5%Sonic 3.6 — 299 ms, 1.5%Falcon 2 — 300 ms, 1.7%Coda — 332 ms, 9.5%S2.1 Pro — 341 ms, 1.5%Grok TTS — 363 ms, 1.6%S1 — 389 ms, 1.8%Lightning v3.1 Pro — 392 ms, 1.5%Daytona TTS — 486 ms, 1.6%Chirp 3 HD — 597 ms, 1.5%Qwen3 TTS Flash Realtime — 718 ms, 3.0%S2.1 Pro Free — 850 ms, 1.5%GPT-4o mini TTS — 2585 ms, 1.8%Flux TTS — 205 ms, 1.7%Aura 2 — 297 ms, 1.8%Aura 2 En — 569 ms, 2.1%
Benchmarked models with their Time to First Audio over the last 30 days.
ModelHostTTFARank
Aura 2 EnCloudflare569 ms28th of 32
Aura 2Deepgram297 ms19th of 32
Flux TTSDeepgram205 ms13th of 32

How fast are Deepgram's STT and TTS models?

Nova 3 measures mean 86 ms time to final segment (5th of 29) among STT systems. Last measured 2026-10-14. Nova 2 measures mean 87 ms time to final segment (6th of 29) among STT systems. Last measured 2026-10-14. Flux Multilingual measures mean 98 ms time to final segment (8th of 29) among STT systems. Last measured 2026-10-14. Flux measures mean 99 ms time to final segment (9th of 29) among STT systems. Last measured 2026-10-14. Flux TTS measures mean 205 ms time to first audio (13th of 32) among TTS systems. Last measured 2026-10-14. Aura 2 measures mean 297 ms time to first audio (19th of 32) among TTS systems. Last measured 2026-10-14. Aura 2 En on Cloudflare measures mean 569 ms time to first audio (28th of 32) among TTS systems. Last measured 2026-10-14.

How accurate are Deepgram's STT and TTS models?

Nova 3 measures 5.3% word error rate (20th of 31) among STT systems. Last measured 2026-10-14. Flux measures 5.8% word error rate (22nd of 31) among STT systems. Last measured 2026-10-14. Nova 2 measures 6.7% word error rate (25th of 31) among STT systems. Last measured 2026-10-14. Flux Multilingual measures 7.5% word error rate (28th of 31) among STT systems. Last measured 2026-10-14. Flux TTS measures 1.7% word error rate (13th of 32) among TTS systems. Last measured 2026-10-14. Aura 2 measures 1.8% word error rate (21st of 32) among TTS systems. Last measured 2026-10-14. Aura 2 En on Cloudflare measures 2.1% word error rate (25th of 32) among TTS systems. Last measured 2026-10-14.

Which Deepgram model is fastest?

Its fastest dated STT result is Nova 3 at mean 86 ms time to final segment (5th of 29) among STT systems, with 5.3% WER. Last measured 2026-10-14. Its fastest dated TTS result is Flux TTS at mean 205 ms time to first audio (13th of 32) among TTS systems, with 1.7% WER. Last measured 2026-10-14.

Limits of this comparison

Coval measures Nova and Flux transcription models and Aura speech output.

  • Coval does not evaluate every Deepgram domain model, language, voice or endpointing configuration.

Official resources

Same datasets, prompts and metric definitions for every model, measured by Coval’s open-source runner. Full methodology on the overview.

Evaluate your own voice agent

Use Coval to test your production configuration, prompts and calls—not only the public benchmark endpoints.

Book a Demo