Voice AI BenchmarksBy

SONIOXSPEECH-TO-TEXT117,665 SAMPLES / 30 DAYSLAST RUN OCT 14, 2026, 06:00 UTC

official resource

STT RT v5 speech-to-text benchmarks

STT RT v5, hosted by Soniox, measures mean 45 ms time to final segment (2nd of 29) and 4.8% word error rate (18th of 31) among STT systems. Results cover the last 30 days. Coval benchmarks hosted endpoints daily using a consistent evaluation methodology. Last measured .

STT RT v5 is Soniox's fifth-generation real-time transcription model.

Time to Final Segment#2 / 29
45ms
Word Error Rate#18 / 31
4.8%
Time to First Token#13 / 27
1518ms

Overview

The Soniox endpoint supports multilingual transcription with diarization, translation, code switching and keyterm controls.

STT RT v5 is tested every day on fixed public audio — clean, accented, noisy, reverberant, far-field, clipped and phone-codec speech — for transcription accuracy and streaming latency.

Technical specifications

Made by
Soniox
Hosted by
Soniox
Source
Official API
Licensing
Proprietary
Deployment
Cloud
Region
US
Features
Code switching, Diarization, Keyterm biasing, Multilingual, Translation, VAD

How STT RT v5 ranks

Full STT dashboard
Time to Final Segment — every measured STT modelMilliseconds · lower is better · 30-day averageEvery STT model ranked on Time to Final Segment, with STT RT v5 highlighted.
  1. #1Qwen3 ASR 1.7b39 ms
  2. #2STT RT v545 ms
  3. #3Qwen3 ASR Fast47 ms
  4. #4STT 166 ms
  5. #5Nova 385 ms
  6. #6Nova 287 ms
  7. #7Parakeet TDT 0.6B v397 ms
  8. #8Flux Multilingual98 ms
  9. #9Flux99 ms
  10. #10Scribe v2 Realtime122 ms
  11. #11Whisper Large v3via Baseten131 ms
  12. #12Universal 3.6 Pro135 ms
Show all 29 models
  1. #13Ink 2148 ms
  2. #14Universal 3.5 Pro168 ms
  3. #15Whisper Large v3via Together AI200 ms
  4. #16Pulse205 ms
  5. #17Grok STT223 ms
  6. #18resonant-1252 ms
  7. #19Default258 ms
  8. #20Linden 1271 ms
  9. #21Gemini 3.5 Transcribe Live317 ms
  10. #22Daytona STT399 ms
  11. #23Voxtral Mini Transcribe Realtime 2602410 ms
  12. #24GPT Realtime Whisper599 ms
  13. #25Solaria 1636 ms
  14. #26GPT-4o mini Transcribe708 ms
  15. #27Chirp 3768 ms
  16. #28GPT-4o Transcribe783 ms
  17. #29Chirp 2908 ms
median of all models · 200 ms
STT models over the last 30 days, ranked on Time to Final Segment.
#ModelHostTTFSWERTTFTSamples
1Qwen3 ASR 1.7bBaseten39 ms946 ms13,669
2STT RT v5Soniox45 ms1518 ms23,533
3Qwen3 ASR FastNari Labs47 ms1752 ms26,652
4STT 1Inworld AI66 ms1211 ms24,749
5Nova 3Deepgram85 ms1431 ms26,002
6Nova 2Deepgram87 ms1417 ms25,960
7Parakeet TDT 0.6B v3Together AI97 ms1228 ms26,461
8Flux MultilingualDeepgram98 ms1165 ms26,007
9FluxDeepgram99 ms1091 ms26,005
10Scribe v2 RealtimeElevenLabs122 ms2158 ms26,640
11Whisper Large v3Baseten131 ms924 ms13,663
12Universal 3.6 ProAssemblyAI135 ms1072 ms24,204
13Ink 2Cartesia148 ms1966 ms26,652
14Universal 3.5 ProAssemblyAI168 ms1025 ms26,654
15Whisper Large v3Together AI200 ms1261 ms26,445
16PulseSmallest205 ms1976 ms26,654
17Grok STTSpaceXAI223 ms—26,576
18resonant-1Reson8252 ms—26,653
19DefaultGradium258 ms1986 ms26,632
20Linden 1Speechmatics271 ms1537 ms26,595
21Gemini 3.5 Transcribe LiveGemini317 ms1609 ms26,567
22Daytona STTGuava399 ms5695 ms5,225
23Voxtral Mini Transcribe Realtime 2602Mistral410 ms1851 ms26,200
24GPT Realtime WhisperOpenAI599 ms1824 ms26,622
25Solaria 1Gladia636 ms1781 ms26,200
26GPT-4o mini TranscribeOpenAI708 ms—26,621
27Chirp 3Google768 ms5956 ms26,638
28GPT-4o TranscribeOpenAI783 ms—26,625
29Chirp 2Google908 ms6098 ms26,650
—Nemotron 3.5 ASR StreamingTogether AI—1544 ms26,542
—Universal StreamingAssemblyAI—1508 ms26,624
Under-sampled models are excluded; tied models share a place. Dotted WER values split into substitutions, deletions and insertions on hover or tap.

Latency vs accuracy

TTFS vs WEREach point is one measured model · STT RT v5 highlighted · 30-day averagesTime to Final Segment against Word Error Rate for every measured STT model, with STT RT v5 highlighted.
Qwen3 ASR 1.7b — 39 ms, 2.9%Qwen3 ASR Fast — 47 ms, 2.6%STT 1 — 66 ms, 7.1%Nova 3 — 85 ms, 5.3%Nova 2 — 87 ms, 6.7%Parakeet TDT 0.6B v3 — 97 ms, 11.1%Flux Multilingual — 98 ms, 7.5%Flux — 99 ms, 5.8%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 — 148 ms, 3.6%Universal 3.5 Pro — 168 ms, 2.4%Whisper Large v3 via Together AI — 200 ms, 7.4%Pulse — 205 ms, 4.6%Grok STT — 223 ms, 4.1%resonant-1 — 252 ms, 2.6%Default — 258 ms, 8.8%Linden 1 — 271 ms, 3.9%Gemini 3.5 Transcribe Live — 317 ms, 3.0%Daytona STT — 399 ms, 4.9%Voxtral Mini Transcribe Realtime 2602 — 410 ms, 5.4%GPT Realtime Whisper — 599 ms, 4.0%Solaria 1 — 636 ms, 6.0%GPT-4o mini Transcribe — 708 ms, 3.9%Chirp 3 — 768 ms, 3.2%GPT-4o Transcribe — 783 ms, 3.8%Chirp 2 — 908 ms, 3.9%STT RT v5 — 45 ms, 4.8%

Where the errors come from

WER compositionSTT RT v5's Word Error Rate split by error type · 30-day averageSTT RT v5's WER split into substitutions, deletions and insertions.
  • STT RT v54.8%
SubstitutionsDeletionsInsertions

Averages and tail latency

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

STT RT v5 latency distributionMilliseconds · shared axis across metrics · last 30 daysSTT RT v5's latency percentiles per metric: p25–p75 band, p50 tick, whisker to max.
  • Time to Final Segmentp50 39 ms · Max 4373 ms
  • Time to First Tokenp50 1535 ms · Max 5483 ms
band p25–p75 · tick p50 · whisker to max with p90 and p95 stops
Average and percentile values per metric, with the number of samples behind each row.
MetricAveragep25p50p75p90p95MaxSamples
Time to Final Segment45 ms32 ms39 ms55 ms63 ms71 ms4373 ms23,533
Word Error Rate4.8%0.0%0.0%8.3%15.8%21.7%100.0%23,533
Time to First Token1518 ms1150 ms1535 ms1727 ms2048 ms2326 ms5483 ms23,533

Last 30 days

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

Time to Final Segment — daily p50Line p50 · band p25–p75 · UTC daysSTT RT v5's daily median Time to Final Segment over the last 30 days.
STT RT v5 · Sep 14: 62 msSTT RT v5 · Sep 18: 55 msSTT RT v5 · Sep 19: 44 msSTT RT v5 · Sep 20: 35 msSTT RT v5 · Sep 21: 47 msSTT RT v5 · Sep 22: 41 msSTT RT v5 · Sep 23: 55 msSTT RT v5 · Sep 24: 41 msSTT RT v5 · Sep 25: 46 msSTT RT v5 · Sep 26: 40 msSTT RT v5 · Sep 27: 35 msSTT RT v5 · Sep 28: 40 msSTT RT v5 · Sep 29: 41 msSTT RT v5 · Sep 30: 39 msSTT RT v5 · Oct 1: 35 msSTT RT v5 · Oct 2: 38 msSTT RT v5 · Oct 3: 35 msSTT RT v5 · Oct 4: 29 msSTT RT v5 · Oct 5: 36 msSTT RT v5 · Oct 6: 42 msSTT RT v5 · Oct 7: 44 msSTT RT v5 · Oct 8: 36 msSTT RT v5 · Oct 9: 45 msSTT RT v5 · Oct 10: 38 msSTT RT v5 · Oct 11: 32 msSTT RT v5 · Oct 12: 44 msSTT RT v5 · Oct 13: 42 msSTT RT v5 · Oct 14: 34 ms
Word Error Rate — daily averageDaily average · UTC daysSTT RT v5's daily Word Error Rate over the last 30 days.
STT RT v5 · Sep 14: 5.2%STT RT v5 · Sep 18: 5.3%STT RT v5 · Sep 19: 5.0%STT RT v5 · Sep 20: 6.9%STT RT v5 · Sep 21: 6.4%STT RT v5 · Sep 22: 5.8%STT RT v5 · Sep 23: 5.3%STT RT v5 · Sep 24: 6.0%STT RT v5 · Sep 25: 6.9%STT RT v5 · Sep 26: 4.7%STT RT v5 · Sep 27: 5.9%STT RT v5 · Sep 28: 7.0%STT RT v5 · Sep 29: 5.3%STT RT v5 · Sep 30: 5.1%STT RT v5 · Oct 1: 5.7%STT RT v5 · Oct 2: 5.1%STT RT v5 · Oct 3: 6.4%STT RT v5 · Oct 4: 5.6%STT RT v5 · Oct 5: 6.9%STT RT v5 · Oct 6: 7.3%STT RT v5 · Oct 7: 4.9%STT RT v5 · Oct 8: 5.5%STT RT v5 · Oct 9: 5.0%STT RT v5 · Oct 10: 5.7%STT RT v5 · Oct 11: 5.5%STT RT v5 · Oct 12: 6.1%STT RT v5 · Oct 13: 5.1%STT RT v5 · Oct 14: 9.0%

Time to Final Segment by dataset

STT RT v5 by test conditionMilliseconds · lower is better · best condition firstSTT RT v5's Time to Final Segment on each benchmark dataset.
  1. WildASR far-field43 ms
  2. WildASR clipping43 ms
  3. WildASR noise gaps44 ms
  4. WildASR clean44 ms
  5. WildASR phone codec44 ms
  6. PipeCat (production)45 ms
  7. WildASR accents46 ms
Time to Final Segment per dataset, with the number of samples behind each figure.
DatasetTTFSSamples
ProductionPipeCat45 ms12,376
AccentsWildASR46 ms1,239
CleanWildASR44 ms4,955
ClippingWildASR43 ms1,241
Far-fieldWildASR43 ms1,242
Noise gapsWildASR44 ms1,239
Phone codecWildASR44 ms1,241

Strongest condition: WildASR far-field at 43 ms · weakest: WildASR accents at 46 ms.

How fast is STT RT v5?

On Soniox, STT RT v5 measures mean 45 ms time to final segment (2nd of 29) and mean 1518 ms time to first token (13th of 27). Last measured 2026-10-14.

How accurate is STT RT v5?

On Soniox, STT RT v5 measures 4.8% word error rate (18th of 31). Last measured 2026-10-14.

Who hosts STT RT v5?

STT RT v5 is created by Soniox and served by Soniox. Coval measures each hosted endpoint separately.

Limits of this comparison

Coval measures it alongside specialist and general-purpose recognizers with the same reference normalization and audio rotation.

  • The benchmark does not score translation or diarization quality, and English-heavy datasets do not establish performance for 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