Voice AI BenchmarksBy

SPEECHMATICSSPEECH-TO-TEXT114,214 SAMPLES / 30 DAYSLAST RUN SEP 15, 2026, 07:30 UTC

official resource

Enhanced speech-to-text benchmarks

Enhanced, hosted by Speechmatics, measures mean 299 ms time to final segment (20th of 28) and 4.3% word error rate (7th of 30) among STT systems. Results cover the last 30 days. Coval benchmarks hosted endpoints daily using a consistent evaluation methodology. Last measured .

Enhanced is Speechmatics' higher-accuracy real-time transcription tier.

Time to Final Segment#20 / 28
299ms
Word Error Rate#7 / 30
4.3%
Time to First Token#12 / 26
1492ms

Overview

Speechmatics exposes Enhanced as a distinct real-time operating point on a service that also supports multilingual transcription, diarization and code switching.

Enhanced 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
Speechmatics
Hosted by
Speechmatics
Source
Official API
Licensing
Proprietary
Deployment
On-prem
Region
Europe
Features
Code switching, Diarization, Keyterm biasing, Multilingual, Translation, VAD

How Enhanced 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 Enhanced highlighted.
  1. #1Qwen3 ASR 1.7b39 ms
  2. #2Qwen3 ASR Fast46 ms
  3. #3STT RT v557 ms
  4. #4STT 165 ms
  5. #5Parakeet TDT 0.6B v381 ms
  6. #6Nova 389 ms
  7. #7Nova 292 ms
  8. #8Flux Multilingual98 ms
  9. #9Flux99 ms
  10. #10Ink 2122 ms
  11. #11Whisper Large v3via Baseten125 ms
  12. #12Scribe v2 Realtime133 ms
  13. #13Universal 3.5 Pro173 ms
  14. #14Grok STT207 ms
  15. #15Defaultvia Speechmatics209 ms
  16. #16Pulse212 ms
  17. #17Defaultvia Gradium246 ms
  18. #18resonant-1264 ms
  19. #19Whisper Large v3via Together AI296 ms
  20. #20Enhanced299 ms
Show all 28 models
  1. #21Gemini 3.5 Transcribe Live305 ms
  2. #22Voxtral Mini Transcribe Realtime 2602400 ms
  3. #23GPT Realtime Whisper551 ms
  4. #24GPT-4o mini Transcribe690 ms
  5. #25GPT-4o Transcribe754 ms
  6. #26Chirp 3776 ms
  7. #27Solaria 1803 ms
  8. #28Chirp 2873 ms
median of all models · 208 ms
STT models over the last 30 days, ranked on Time to Final Segment.
#ModelHostTTFSWERTTFTSamples
1Qwen3 ASR 1.7bBaseten39 ms931 ms1,256
2Qwen3 ASR FastNari46 ms1748 ms4,391
3STT RT v5Soniox57 ms1529 ms22,468
4STT 1Inworld AI65 ms1400 ms22,776
5Parakeet TDT 0.6B v3Together AI81 ms1215 ms22,428
6Nova 3Deepgram89 ms1418 ms22,795
7Nova 2Deepgram92 ms1419 ms22,673
8Flux MultilingualDeepgram98 ms1163 ms12,191
9FluxDeepgram99 ms1076 ms12,183
10Ink 2Cartesia122 ms1827 ms22,802
11Whisper Large v3Baseten125 ms912 ms1,252
12Scribe v2 RealtimeElevenLabs133 ms2175 ms22,806
13Universal 3.5 ProAssemblyAI173 ms1040 ms20,193
14Grok STTxAI207 ms—22,793
15DefaultSpeechmatics209 ms1428 ms22,816
16PulseSmallest212 ms2011 ms22,796
17DefaultGradium246 ms1980 ms22,779
18resonant-1Reson8264 ms—22,800
19Whisper Large v3Together AI296 ms1338 ms22,685
20EnhancedSpeechmatics299 ms1492 ms22,816
21Gemini 3.5 Transcribe LiveGemini305 ms1660 ms16,343
22Voxtral Mini Transcribe Realtime 2602Mistral400 ms1850 ms22,526
23GPT Realtime WhisperOpenAI551 ms1814 ms22,752
24GPT-4o mini TranscribeOpenAI690 ms—22,770
25GPT-4o TranscribeOpenAI754 ms—22,771
26Chirp 3Google776 ms5974 ms22,815
27Solaria 1Gladia803 ms1801 ms22,379
28Chirp 2Google873 ms6072 ms22,729
—Nemotron 3.5 ASR StreamingTogether AI—1548 ms22,687
—Universal StreamingAssemblyAI—1513 ms20,175
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: 7th of 30 on Word Error Rate.

Latency vs accuracy

TTFS vs WEREach point is one measured model · Enhanced highlighted · 30-day averagesTime to Final Segment against Word Error Rate for every measured STT model, with Enhanced highlighted.
Qwen3 ASR 1.7b — 39 ms, 4.1%Qwen3 ASR Fast — 46 ms, 3.2%STT RT v5 — 57 ms, 5.7%STT 1 — 65 ms, 4.4%Parakeet TDT 0.6B v3 — 81 ms, 11.2%Nova 3 — 89 ms, 6.2%Nova 2 — 92 ms, 7.9%Flux Multilingual — 98 ms, 8.8%Flux — 99 ms, 6.6%Ink 2 — 122 ms, 4.8%Whisper Large v3 via Baseten — 125 ms, 5.6%Scribe v2 Realtime — 133 ms, 5.5%Universal 3.5 Pro — 173 ms, 3.1%Grok STT — 207 ms, 4.6%Default via Speechmatics — 209 ms, 5.4%Pulse — 212 ms, 5.3%Default via Gradium — 246 ms, 10.0%resonant-1 — 264 ms, 3.4%Whisper Large v3 via Together AI — 296 ms, 8.5%Gemini 3.5 Transcribe Live — 305 ms, 3.9%Voxtral Mini Transcribe Realtime 2602 — 400 ms, 6.2%GPT Realtime Whisper — 551 ms, 5.1%GPT-4o mini Transcribe — 690 ms, 4.9%GPT-4o Transcribe — 754 ms, 4.6%Chirp 3 — 776 ms, 4.1%Solaria 1 — 803 ms, 7.1%Chirp 2 — 873 ms, 4.9%Enhanced — 299 ms, 4.3%

Where the errors come from

WER compositionEnhanced's Word Error Rate split by error type · 30-day averageEnhanced's WER split into substitutions, deletions and insertions.
  • Enhanced4.3%
SubstitutionsDeletionsInsertions

Averages and tail latency

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

Enhanced latency distributionMilliseconds · shared axis across metrics · last 30 daysEnhanced's latency percentiles per metric: p25–p75 band, p50 tick, whisker to p99.
  • Time to Final Segmentp50 288 ms · p99 520 ms
  • Time to First Tokenp50 1517 ms · p99 3293 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 Final Segment299 ms255 ms288 ms329 ms377 ms413 ms520 ms22,816
Word Error Rate4.3%0.0%0.0%6.3%12.5%18.2%33.3%22,846
Time to First Token1492 ms1125 ms1517 ms1602 ms1978 ms2259 ms3293 ms22,846

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 daysEnhanced's daily median Time to Final Segment over the last 30 days.
Enhanced · Sep 1: 303 msEnhanced · Sep 2: 297 msEnhanced · Sep 3: 305 msEnhanced · Sep 4: 304 msEnhanced · Sep 5: 266 msEnhanced · Sep 6: 259 msEnhanced · Sep 7: 268 msEnhanced · Sep 8: 302 msEnhanced · Sep 9: 290 msEnhanced · Sep 10: 279 msEnhanced · Sep 11: 279 msEnhanced · Sep 12: 249 msEnhanced · Sep 13: 241 msEnhanced · Sep 14: 273 msEnhanced · Sep 15: 263 ms
Word Error Rate — daily averageDaily average · UTC daysEnhanced's daily Word Error Rate over the last 30 days.
Enhanced · Sep 1: 9.1%Enhanced · Sep 2: 4.9%Enhanced · Sep 3: 5.4%Enhanced · Sep 4: 4.8%Enhanced · Sep 5: 4.8%Enhanced · Sep 6: 5.0%Enhanced · Sep 7: 5.0%Enhanced · Sep 8: 4.9%Enhanced · Sep 9: 5.9%Enhanced · Sep 10: 4.5%Enhanced · Sep 11: 4.3%Enhanced · Sep 12: 4.3%Enhanced · Sep 13: 4.9%Enhanced · Sep 14: 5.3%Enhanced · Sep 15: 4.9%

Time to Final Segment by dataset

Enhanced by test conditionMilliseconds · lower is better · best condition firstEnhanced's Time to Final Segment on each benchmark dataset.
  1. PipeCat (production)295 ms
  2. WildASR clean298 ms
  3. WildASR phone codec299 ms
  4. WildASR far-field303 ms
  5. WildASR noise gaps305 ms
  6. WildASR reverb305 ms
  7. WildASR clipping307 ms
  8. WildASR accents324 ms
Time to Final Segment per dataset, with the number of samples behind each figure.
DatasetTTFSSamples
LibriSpeech371 ms1
ProductionPipeCat295 ms11,489
AccentsWildASR324 ms1,114
CleanWildASR298 ms4,567
ClippingWildASR307 ms1,137
Far-fieldWildASR303 ms1,140
Noise gapsWildASR305 ms1,140
Phone codecWildASR299 ms1,135
ReverbWildASR305 ms1,093

Strongest condition: PipeCat (production) at 295 ms · weakest: LibriSpeech at 371 ms.

How fast is Enhanced?

On Speechmatics, Enhanced measures mean 299 ms time to final segment (20th of 28) and mean 1492 ms time to first token (12th of 26). Last measured 2026-09-15.

How accurate is Enhanced?

On Speechmatics, Enhanced measures 4.3% word error rate (7th of 30). Last measured 2026-09-15.

Who hosts Enhanced?

Enhanced is created by Speechmatics and served by Speechmatics. Coval measures each hosted endpoint separately.

Limits of this comparison

Enhanced remains separate from the provider's default tier so the extra recognition mode is not blended into a company-level average.

  • The benchmark measures transcription outcomes and timing, not translation quality or every language available through Speechmatics.

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