Voice AI BenchmarksBy

GOOGLESPEECH-TO-TEXT146,759 SAMPLES / 30 DAYSLAST RUN AUG 28, 2026, 21:30 UTC

official resource

Chirp 3 speech-to-text benchmarks

Chirp 3 is Google's newer Chirp speech recognition model for Cloud Speech-to-Text.

Time to Final Segment#24 / 24
813ms
Word Error Rate#3 / 28
4.1%
Time to First Token#23 / 24
5999ms

Overview

Chirp 3 is Google's latest multilingual foundation speech model, with voice-activity and keyterm support on the tested endpoint.

Chirp 3 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 3 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 3 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
  24. #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: 3rd of 28 on Word Error Rate.

Latency vs accuracy

TTFS vs WEREach point is one measured model · Chirp 3 highlighted · 30-day averagesTime to Final Segment against Word Error Rate for every measured STT model, with Chirp 3 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 2 — 811 ms, 5.0%Chirp 3 — 813 ms, 4.1%

Where the errors come from

WER compositionChirp 3's Word Error Rate split by error type · 30-day averageChirp 3's WER split into substitutions, deletions and insertions.
  • Chirp 34.1%
SubstitutionsDeletionsInsertions

Averages and tail latency

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

Chirp 3 latency distributionMilliseconds · shared axis across metrics · last 30 daysChirp 3's latency percentiles per metric: p25–p75 band, p50 tick, whisker to p99.
  • Time to Final Segmentp50 789 ms · p99 1603 ms
  • Time to First Tokenp50 6370 ms · p99 9047 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 Segment813 ms695 ms789 ms875 ms962 ms1050 ms1603 ms29,319
Word Error Rate4.1%0.0%0.0%5.3%12.0%18.8%41.7%29,360
Time to First Token5999 ms5693 ms6370 ms6596 ms6951 ms7341 ms9047 ms29,360

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 3's daily median Time to Final Segment over the last 30 days.
Chirp 3 · Jul 30: 806 msChirp 3 · Jul 31: 738 msChirp 3 · Aug 1: 787 msChirp 3 · Aug 2: 666 msChirp 3 · Aug 3: 780 msChirp 3 · Aug 4: 839 msChirp 3 · Aug 5: 912 msChirp 3 · Aug 6: 828 msChirp 3 · Aug 7: 777 msChirp 3 · Aug 8: 817 msChirp 3 · Aug 9: 785 msChirp 3 · Aug 10: 778 msChirp 3 · Aug 11: 847 msChirp 3 · Aug 12: 809 msChirp 3 · Aug 13: 806 msChirp 3 · Aug 14: 794 msChirp 3 · Aug 15: 796 msChirp 3 · Aug 16: 832 msChirp 3 · Aug 17: 782 msChirp 3 · Aug 18: 802 msChirp 3 · Aug 19: 805 msChirp 3 · Aug 20: 886 msChirp 3 · Aug 21: 779 msChirp 3 · Aug 22: 797 msChirp 3 · Aug 23: 703 msChirp 3 · Aug 24: 675 msChirp 3 · Aug 25: 842 msChirp 3 · Aug 26: 707 msChirp 3 · Aug 27: 719 msChirp 3 · Aug 28: 765 ms
Word Error Rate — daily averageDaily average · UTC daysChirp 3's daily Word Error Rate over the last 30 days.
Chirp 3 · Jul 30: 4.2%Chirp 3 · Jul 31: 5.3%Chirp 3 · Aug 1: 4.5%Chirp 3 · Aug 2: 6.1%Chirp 3 · Aug 3: 4.5%Chirp 3 · Aug 4: 5.6%Chirp 3 · Aug 5: 6.4%Chirp 3 · Aug 6: 4.3%Chirp 3 · Aug 7: 6.7%Chirp 3 · Aug 8: 6.3%Chirp 3 · Aug 9: 5.6%Chirp 3 · Aug 10: 5.7%Chirp 3 · Aug 11: 5.8%Chirp 3 · Aug 12: 3.1%Chirp 3 · Aug 13: 5.3%Chirp 3 · Aug 14: 5.7%Chirp 3 · Aug 15: 4.8%Chirp 3 · Aug 16: 5.8%Chirp 3 · Aug 17: 4.6%Chirp 3 · Aug 18: 5.8%Chirp 3 · Aug 19: 5.9%Chirp 3 · Aug 20: 4.9%Chirp 3 · Aug 21: 3.8%Chirp 3 · Aug 22: 6.8%Chirp 3 · Aug 23: 5.7%Chirp 3 · Aug 24: 5.5%Chirp 3 · Aug 25: 4.4%Chirp 3 · Aug 26: 3.7%Chirp 3 · Aug 27: 6.0%Chirp 3 · Aug 28: 6.1%

Time to Final Segment by dataset

Chirp 3 by test conditionMilliseconds · lower is better · best condition firstChirp 3's Time to Final Segment on each benchmark dataset.
  1. WildASR accents699 ms
  2. WildASR clipping778 ms
  3. WildASR far-field790 ms
  4. WildASR reverb794 ms
  5. WildASR noise gaps798 ms
  6. WildASR clean817 ms
  7. PipeCat (production)829 ms
  8. WildASR phone codec833 ms
Time to Final Segment per dataset, with the number of samples behind each figure.
DatasetTTFSSamples
LibriSpeech850 ms1
ProductionPipeCat829 ms14,935
AccentsWildASR699 ms1,438
CleanWildASR817 ms5,780
ClippingWildASR778 ms1,444
Far-fieldWildASR790 ms1,441
Noise gapsWildASR798 ms1,441
Phone codecWildASR833 ms1,441
ReverbWildASR794 ms1,398

Strongest condition: WildASR accents at 699 ms · weakest: LibriSpeech at 850 ms.

Limits of this comparison

Coval reports transcript accuracy, time to first partial transcript and time to final transcript separately.

  • One benchmark configuration cannot represent every Chirp 3 language, Cloud region or recognizer option.

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