Voice AI BenchmarksBy

GOOGLESPEECH-TO-TEXT146,349 SAMPLES / 30 DAYSLAST RUN AUG 28, 2026, 21:30 UTC

official resource

Chirp 2 speech-to-text benchmarks

Chirp 2 is a Google Cloud Speech-to-Text model in the Chirp family.

Time to Final Segment#23 / 24
811ms
Word Error Rate#9 / 28
5%
Time to First Token#24 / 24
6000ms

Overview

Chirp 2 is a multilingual recognizer served through Cloud Speech-to-Text V2; the tested endpoint includes voice-activity and phrase-biasing controls.

Chirp 2 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
Google
Hosted by
Google
Source
Official API
Licensing
Proprietary
Deployment
On-prem
Region
US
Features
Multilingual, VAD, Keyterm biasing

How Chirp 2 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 Chirp 2 highlighted.
  1. #1STT RT v564 ms
  2. #2Parakeet TDT 0.6B v370 ms
  3. #3STT 183 ms
  4. #4Nova 399 ms
  5. #5Nova 2101 ms
  6. #6Ink 2108 ms
  7. #7Scribe v2 Realtime120 ms
  8. #8Universal 3.5 Pro146 ms
  9. #9Defaultvia Azure155 ms
  10. #10Whisper Large v3180 ms
  11. #11Velma 2 STT Streaming191 ms
  12. #12Grok STT199 ms
  13. #13Pulse205 ms
  14. #14Defaultvia Speechmatics223 ms
  15. #15Defaultvia Gradium249 ms
  16. #16resonant-1288 ms
  17. #17Enhanced341 ms
  18. #18Voxtral Mini Transcribe Realtime 2602356 ms
  19. #19GPT Realtime Whisper559 ms
  20. #20GPT-4o mini Transcribe623 ms
  21. #21Solaria 1680 ms
  22. #22GPT-4o Transcribe742 ms
  23. #23Chirp 2811 ms
Show all 24 models
  1. #24Chirp 3813 ms
median of all models · 202 ms
STT models over the last 30 days, ranked on Time to Final Segment.
#ModelHostTTFSWERTTFTSamples
1STT RT v5Soniox64 ms1533 ms29,322
2Parakeet TDT 0.6B v3Together AI70 ms1210 ms29,161
3STT 1Inworld AI83 ms1468 ms29,280
4Nova 3Deepgram99 ms1434 ms29,290
5Nova 2Deepgram101 ms1436 ms29,145
6Ink 2Cartesia108 ms1812 ms29,288
7Scribe v2 RealtimeElevenLabs120 ms2157 ms29,313
8Universal 3.5 ProAssemblyAI146 ms1030 ms29,280
9DefaultAzure155 ms1792 ms6,682
10Whisper Large v3Together AI180 ms1241 ms29,171
11Velma 2 STT StreamingModulate191 ms1576 ms17,895
12Grok STTxAI199 ms—29,322
13PulseSmallest205 ms2072 ms29,313
14DefaultSpeechmatics223 ms1473 ms29,323
15DefaultGradium249 ms1979 ms28,235
16resonant-1Reson8288 ms—17,066
17EnhancedSpeechmatics341 ms1536 ms29,322
18Voxtral Mini Transcribe Realtime 2602Mistral356 ms1828 ms29,175
19GPT Realtime WhisperOpenAI559 ms1823 ms29,286
20GPT-4o mini TranscribeOpenAI623 ms—29,318
21Solaria 1Gladia680 ms1703 ms26,356
22GPT-4o TranscribeOpenAI742 ms—29,318
23Chirp 2Google811 ms6000 ms29,237
24Chirp 3Google813 ms5999 ms29,319
—FluxDeepgram—1089 ms29,354
—Flux MultilingualDeepgram—1175 ms29,344
—Nemotron 3.5 ASR StreamingTogether AI—1540 ms29,193
—Universal StreamingAssemblyAI—1510 ms29,297
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: 9th of 28 on Word Error Rate.

Latency vs accuracy

TTFS vs WEREach point is one measured model · Chirp 2 highlighted · 30-day averagesTime to Final Segment against Word Error Rate for every measured STT model, with Chirp 2 highlighted.
STT RT v5 — 64 ms, 5.8%Parakeet TDT 0.6B v3 — 70 ms, 11.0%STT 1 — 83 ms, 4.8%Nova 3 — 99 ms, 6.3%Nova 2 — 101 ms, 7.9%Ink 2 — 108 ms, 4.9%Scribe v2 Realtime — 120 ms, 5.5%Universal 3.5 Pro — 146 ms, 3.2%Default via Azure — 155 ms, 5.3%Whisper Large v3 — 180 ms, 8.2%Velma 2 STT Streaming — 191 ms, 6.4%Grok STT — 199 ms, 4.8%Pulse — 205 ms, 5.3%Default via Speechmatics — 223 ms, 5.5%Default via Gradium — 249 ms, 10.0%resonant-1 — 288 ms, 3.4%Enhanced — 341 ms, 4.3%Voxtral Mini Transcribe Realtime 2602 — 356 ms, 6.2%GPT Realtime Whisper — 559 ms, 5.2%GPT-4o mini Transcribe — 623 ms, 5.1%Solaria 1 — 680 ms, 7.1%GPT-4o Transcribe — 742 ms, 4.7%Chirp 3 — 813 ms, 4.1%Chirp 2 — 811 ms, 5.0%

Where the errors come from

WER compositionChirp 2's Word Error Rate split by error type · 30-day averageChirp 2's WER split into substitutions, deletions and insertions.
  • Chirp 25.0%
SubstitutionsDeletionsInsertions

Averages and tail latency

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

Chirp 2 latency distributionMilliseconds · shared axis across metrics · last 30 daysChirp 2's latency percentiles per metric: p25–p75 band, p50 tick, whisker to p99.
  • Time to Final Segmentp50 767 ms · p99 1586 ms
  • Time to First Tokenp50 6385 ms · p99 8905 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 Segment811 ms681 ms767 ms854 ms1076 ms1226 ms1586 ms29,237
Word Error Rate5.0%0.0%0.0%7.1%14.3%20.0%40.9%29,278
Time to First Token6000 ms5678 ms6385 ms6595 ms6991 ms7332 ms8905 ms29,278

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 daysChirp 2's daily median Time to Final Segment over the last 30 days.
Chirp 2 · Jul 30: 768 msChirp 2 · Jul 31: 677 msChirp 2 · Aug 1: 762 msChirp 2 · Aug 2: 685 msChirp 2 · Aug 3: 790 msChirp 2 · Aug 4: 731 msChirp 2 · Aug 5: 735 msChirp 2 · Aug 6: 760 msChirp 2 · Aug 7: 737 msChirp 2 · Aug 8: 783 msChirp 2 · Aug 9: 732 msChirp 2 · Aug 10: 722 msChirp 2 · Aug 11: 760 msChirp 2 · Aug 12: 750 msChirp 2 · Aug 13: 759 msChirp 2 · Aug 14: 743 msChirp 2 · Aug 15: 810 msChirp 2 · Aug 16: 753 msChirp 2 · Aug 17: 766 msChirp 2 · Aug 18: 781 msChirp 2 · Aug 19: 808 msChirp 2 · Aug 20: 802 msChirp 2 · Aug 21: 774 msChirp 2 · Aug 22: 796 msChirp 2 · Aug 23: 731 msChirp 2 · Aug 24: 766 msChirp 2 · Aug 25: 783 msChirp 2 · Aug 26: 815 msChirp 2 · Aug 27: 773 msChirp 2 · Aug 28: 828 ms
Word Error Rate — daily averageDaily average · UTC daysChirp 2's daily Word Error Rate over the last 30 days.
Chirp 2 · Jul 30: 6.4%Chirp 2 · Jul 31: 6.5%Chirp 2 · Aug 1: 5.3%Chirp 2 · Aug 2: 5.7%Chirp 2 · Aug 3: 5.2%Chirp 2 · Aug 4: 6.9%Chirp 2 · Aug 5: 6.1%Chirp 2 · Aug 6: 6.1%Chirp 2 · Aug 7: 5.9%Chirp 2 · Aug 8: 6.7%Chirp 2 · Aug 9: 6.5%Chirp 2 · Aug 10: 5.4%Chirp 2 · Aug 11: 6.2%Chirp 2 · Aug 12: 6.2%Chirp 2 · Aug 13: 5.3%Chirp 2 · Aug 14: 6.8%Chirp 2 · Aug 15: 4.7%Chirp 2 · Aug 16: 6.7%Chirp 2 · Aug 17: 5.3%Chirp 2 · Aug 18: 7.3%Chirp 2 · Aug 19: 6.3%Chirp 2 · Aug 20: 5.1%Chirp 2 · Aug 21: 6.2%Chirp 2 · Aug 22: 5.7%Chirp 2 · Aug 23: 5.7%Chirp 2 · Aug 24: 5.1%Chirp 2 · Aug 25: 4.7%Chirp 2 · Aug 26: 5.1%Chirp 2 · Aug 27: 5.9%Chirp 2 · Aug 28: 5.9%

Time to Final Segment by dataset

Chirp 2 by test conditionMilliseconds · lower is better · best condition firstChirp 2's Time to Final Segment on each benchmark dataset.
  1. WildASR accents704 ms
  2. WildASR reverb783 ms
  3. WildASR far-field802 ms
  4. WildASR clipping803 ms
  5. WildASR clean807 ms
  6. WildASR phone codec818 ms
  7. WildASR noise gaps823 ms
  8. PipeCat (production)826 ms
Time to Final Segment per dataset, with the number of samples behind each figure.
DatasetTTFSSamples
LibriSpeech930 ms1
ProductionPipeCat826 ms14,891
AccentsWildASR704 ms1,435
CleanWildASR807 ms5,760
ClippingWildASR803 ms1,440
Far-fieldWildASR802 ms1,441
Noise gapsWildASR823 ms1,439
Phone codecWildASR818 ms1,437
ReverbWildASR783 ms1,393

Strongest condition: WildASR accents at 704 ms · weakest: LibriSpeech at 930 ms.

Limits of this comparison

Coval keeps it distinct from Chirp 3 so a generation change can be evaluated on the same audio and scoring rules.

  • The results describe Coval's configured US endpoint; Cloud region, recognizer settings and adaptation choices can change production behavior.

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