Voice AI BenchmarksBy

SONIOXTEXT-TO-SPEECH44,916 SAMPLES / 30 DAYSLAST RUN SEP 14, 2026, 22:30 UTC

official resource

TTS RT v1 text-to-speech benchmarks

TTS RT v1, hosted by Soniox, measures mean 245 ms time to first audio (10th of 28) and 3.9% word error rate (1st of 28) among TTS systems. Results cover the last 30 days. Coval benchmarks hosted endpoints daily using a consistent evaluation methodology. Last measured .

TTS RT v1 is Soniox's first versioned real-time synthesis endpoint.

Time to First Audio#10 / 28
245ms
TTFA Network Roundtrip#13 / 28
170ms
TTFA Leading Silence#14 / 28
75ms
Word Error Rate#1 / 28
3.9%

Overview

Soniox provides the endpoint for streaming multilingual synthesis with voice cloning.

TTS RT v1 is tested every day on a fixed set of text prompts, measuring how quickly audible speech starts and how intelligible the result is.

Technical specifications

Made by
Soniox
Hosted by
Soniox
Source
Official API
Licensing
Proprietary
Deployment
Cloud
Region
US
Features
Multilingual, Voice cloning

How TTS RT v1 ranks

Full TTS dashboard
Time to First Audio — every measured TTS modelMilliseconds · lower is better · 30-day averageEvery TTS model ranked on Time to First Audio, with TTS RT v1 highlighted.
  1. #1vui66 ms
  2. #2Qwen3 TTS Fast71 ms
  3. #3TTS Flash 291 ms
  4. #4Qwen3 TTS 1.7b106 ms
  5. #5Palabra TTS v1115 ms
  6. #6TTS 2181 ms
  7. #7Flash v2.5194 ms
  8. #8Default235 ms
  9. #9TTS Rt v2245 ms
  10. #10TTS RT v1245 ms
  11. #11Mist v3258 ms
  12. #12Sonic 3.5277 ms
Show all 28 models
  1. #13Phantom Z 3.4 conversational286 ms
  2. #14Aura 2308 ms
  3. #15Coda312 ms
  4. #16S2.1 Pro335 ms
  5. #17Eleven v3 Conversational348 ms
  6. #18Lightning v3.1 Pro362 ms
  7. #19S1379 ms
  8. #20Grok TTS397 ms
  9. #21Sonic 3.6423 ms
  10. #22Simba 3.2451 ms
  11. #23Simba 3.0472 ms
  12. #24Chirp 3 HD535 ms
  13. #25Falcon 2545 ms
  14. #26Qwen3 TTS Flash Realtime753 ms
  15. #27S2.1 Pro Free964 ms
  16. #28GPT-4o mini TTS1013 ms
median of all models · 310 ms
TTS models over the last 30 days, ranked on Time to First Audio.
#ModelHostTTFAWERSamples
1vuiFluxions66 ms9,470
2Qwen3 TTS FastNari71 ms2,210
3TTS Flash 2Inworld AI91 ms11,421
4Qwen3 TTS 1.7bBaseten106 ms420
5Palabra TTS v1Palabra115 ms11,405
6TTS 2Inworld AI181 ms11,421
7Flash v2.5ElevenLabs194 ms9,138
8DefaultGradium235 ms9,147
9TTS Rt v2Soniox245 ms11,232
10TTS RT v1Soniox245 ms11,234
11Mist v3Rime258 ms11,407
12Sonic 3.5Cartesia277 ms11,420
13Phantom Z 3.4 conversationalDeepdub286 ms11,416
14Aura 2Deepgram308 ms11,406
15CodaRime312 ms11,412
16S2.1 ProFish Audio335 ms11,420
17Eleven v3 ConversationalElevenLabs348 ms11,409
18Lightning v3.1 ProSmallest362 ms11,421
19S1Fish Audio379 ms11,413
20Grok TTSxAI397 ms11,406
21Sonic 3.6Cartesia423 ms8,230
22Simba 3.2Speechify451 ms11,413
23Simba 3.0Speechify472 ms11,416
24Chirp 3 HDGoogle535 ms11,420
25Falcon 2Murf545 ms11,417
26Qwen3 TTS Flash RealtimeAlibaba753 ms11,246
27S2.1 Pro FreeFish Audio964 ms11,386
28GPT-4o mini TTSOpenAI1013 ms11,390
Under-sampled models are excluded; tied models share a place. Dotted WER values split into substitutions, deletions and insertions on hover or tap.

Highest relative placement: 1st of 28 on Word Error Rate.

Latency vs accuracy

TTFA vs WEREach point is one measured model · TTS RT v1 highlighted · 30-day averagesTime to First Audio against Word Error Rate for every measured TTS model, with TTS RT v1 highlighted.
vui — 66 ms, 5.4%Qwen3 TTS Fast — 71 ms, 4.2%TTS Flash 2 — 91 ms, 5.4%Qwen3 TTS 1.7b — 106 ms, 5.5%Palabra TTS v1 — 115 ms, 5.8%TTS 2 — 181 ms, 4.6%Flash v2.5 — 194 ms, 6.7%Default — 235 ms, 5.3%TTS Rt v2 — 245 ms, 4.2%Mist v3 — 258 ms, 5.6%Sonic 3.5 — 277 ms, 5.9%Phantom Z 3.4 conversational — 286 ms, 6.0%Aura 2 — 308 ms, 5.2%Coda — 312 ms, 5.0%S2.1 Pro — 335 ms, 4.8%Eleven v3 Conversational — 348 ms, 4.4%Lightning v3.1 Pro — 362 ms, 4.4%S1 — 379 ms, 4.9%Grok TTS — 397 ms, 4.9%Sonic 3.6 — 423 ms, 5.3%Simba 3.2 — 451 ms, 4.4%Simba 3.0 — 472 ms, 5.2%Chirp 3 HD — 535 ms, 5.2%Falcon 2 — 545 ms, 5.2%Qwen3 TTS Flash Realtime — 753 ms, 8.9%S2.1 Pro Free — 964 ms, 4.7%GPT-4o mini TTS — 1013 ms, 4.8%TTS RT v1 — 245 ms, 3.9%

Where the errors come from

WER compositionTTS RT v1's Word Error Rate split by error type · 30-day averageTTS RT v1's WER split into substitutions, deletions and insertions.
  • TTS RT v13.9%
SubstitutionsDeletionsInsertions

Averages and tail latency

Averages hide slow outliers — these are the distributions behind each figure.

TTS RT v1 latency distributionMilliseconds · shared axis across metrics · last 30 daysTTS RT v1's latency percentiles per metric: p25–p75 band, p50 tick, whisker to p99.
  • Time to First Audiop50 235 ms · p99 331 ms
  • TTFA Network Roundtripp50 155 ms · p99 246 ms
  • TTFA Leading Silencep50 76 ms · p99 110 ms
band p25–p75 · tick p50 · whisker to p99 with p90 and p95 stops
Average and percentile values per metric, with the number of samples behind each row.
MetricAveragep25p50p75p90p95p99Samples
Time to First Audio245 ms216 ms235 ms274 ms296 ms306 ms331 ms11,234
TTFA Network Roundtrip170 ms142 ms155 ms201 ms215 ms222 ms246 ms11,234
TTFA Leading Silence75 ms65 ms76 ms86 ms93 ms97 ms110 ms11,234
Word Error Rate3.9%0.0%0.0%5.9%20.0%20.0%28.6%11,214

Last 30 days

Daily medians from the same measurement runs · gaps are days without qualifying runs.

Time to First Audio — daily p50Line p50 · band p25–p75 · UTC daysTTS RT v1's daily median Time to First Audio over the last 30 days.
TTS RT v1 · Sep 1: 227 msTTS RT v1 · Sep 2: 242 msTTS RT v1 · Sep 3: 234 msTTS RT v1 · Sep 4: 233 msTTS RT v1 · Sep 5: 222 msTTS RT v1 · Sep 6: 233 msTTS RT v1 · Sep 7: 240 msTTS RT v1 · Sep 8: 263 msTTS RT v1 · Sep 9: 259 msTTS RT v1 · Sep 10: 247 msTTS RT v1 · Sep 11: 270 msTTS RT v1 · Sep 12: 263 msTTS RT v1 · Sep 13: 260 msTTS RT v1 · Sep 14: 253 ms
Word Error Rate — daily averageDaily average · UTC daysTTS RT v1's daily Word Error Rate over the last 30 days.
TTS RT v1 · Sep 1: 2.6%TTS RT v1 · Sep 2: 3.4%TTS RT v1 · Sep 3: 3.8%TTS RT v1 · Sep 4: 3.9%TTS RT v1 · Sep 5: 4.5%TTS RT v1 · Sep 6: 5.1%TTS RT v1 · Sep 7: 3.8%TTS RT v1 · Sep 8: 4.4%TTS RT v1 · Sep 9: 4.4%TTS RT v1 · Sep 10: 4.5%TTS RT v1 · Sep 11: 4.4%TTS RT v1 · Sep 12: 3.9%TTS RT v1 · Sep 13: 4.5%TTS RT v1 · Sep 14: 4.4%

Time to First Audio by dataset

TTS RT v1 by test conditionMilliseconds · lower is better · best condition firstTTS RT v1's Time to First Audio on each benchmark dataset.
  1. Text prompts245 ms
Time to First Audio per dataset, with the number of samples behind each figure.
DatasetTTFASamples
Text prompts245 ms11,234

Strongest condition: Text prompts at 245 ms · weakest: Text prompts at 245 ms.

How fast is TTS RT v1?

On Soniox, TTS RT v1 measures mean 245 ms time to first audio (10th of 28). Last measured 2026-09-14.

How accurate is TTS RT v1?

On Soniox, TTS RT v1 measures 3.9% word error rate (1st of 28). Last measured 2026-09-14.

Who hosts TTS RT v1?

TTS RT v1 is created by Soniox and served by Soniox. Coval measures each hosted endpoint separately.

Limits of this comparison

RT v1 remains measured separately from v2 using the same prompts.

  • The benchmark uses one voice and does not rate clone fidelity or every supported language.

Official sources

Results are re-measured daily using fixed datasets and reported over a rolling 30-day window. 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