Voice AI BenchmarksBy

XAISPEECH-TO-TEXT91,443 SAMPLES / 30 DAYSLAST RUN SEP 15, 2026, 08:00 UTC

official resource

Grok STT speech-to-text benchmarks

Grok STT, hosted by xAI, measures mean 207 ms time to final segment (14th of 28) and 4.6% word error rate (10th of 30) among STT systems. Results cover the last 30 days. Coval benchmarks hosted endpoints daily using a consistent evaluation methodology. Last measured .

Grok STT is xAI's managed streaming speech-to-text offering.

Time to Final Segment#14 / 28
207ms
Word Error Rate#10 / 30
4.6%

Overview

The tested endpoint supports multilingual transcription, diarization and keyterm biasing.

Grok STT 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
xAI
Hosted by
xAI
Source
Official API
Licensing
Proprietary
Deployment
Cloud
Region
US
Features
Diarization, Keyterm biasing, Multilingual, VAD

How Grok STT 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 Grok STT 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
Show all 28 models
  1. #15Defaultvia Speechmatics209 ms
  2. #16Pulse212 ms
  3. #17Defaultvia Gradium246 ms
  4. #18resonant-1264 ms
  5. #19Whisper Large v3via Together AI296 ms
  6. #20Enhanced299 ms
  7. #21Gemini 3.5 Transcribe Live305 ms
  8. #22Voxtral Mini Transcribe Realtime 2602400 ms
  9. #23GPT Realtime Whisper551 ms
  10. #24GPT-4o mini Transcribe690 ms
  11. #25GPT-4o Transcribe754 ms
  12. #26Chirp 3776 ms
  13. #27Solaria 1805 ms
  14. #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 ms4,431
3STT RT v5Soniox57 ms1529 ms22,468
4STT 1Inworld AI65 ms1400 ms22,816
5Parakeet TDT 0.6B v3Together AI81 ms1215 ms22,468
6Nova 3Deepgram89 ms1418 ms22,835
7Nova 2Deepgram92 ms1419 ms22,713
8Flux MultilingualDeepgram98 ms1163 ms12,231
9FluxDeepgram99 ms1076 ms12,223
10Ink 2Cartesia122 ms1827 ms22,842
11Whisper Large v3Baseten125 ms912 ms1,252
12Scribe v2 RealtimeElevenLabs133 ms2175 ms22,846
13Universal 3.5 ProAssemblyAI173 ms1040 ms20,233
14Grok STTxAI207 ms—22,833
15DefaultSpeechmatics209 ms1428 ms22,856
16PulseSmallest212 ms2011 ms22,836
17DefaultGradium246 ms1980 ms22,819
18resonant-1Reson8264 ms—22,840
19Whisper Large v3Together AI296 ms1338 ms22,724
20EnhancedSpeechmatics299 ms1491 ms22,856
21Gemini 3.5 Transcribe LiveGemini305 ms1671 ms16,383
22Voxtral Mini Transcribe Realtime 2602Mistral400 ms1850 ms22,564
23GPT Realtime WhisperOpenAI551 ms1814 ms22,792
24GPT-4o mini TranscribeOpenAI690 ms—22,810
25GPT-4o TranscribeOpenAI754 ms—22,811
26Chirp 3Google776 ms5974 ms22,855
27Solaria 1Gladia805 ms1803 ms22,419
28Chirp 2Google873 ms6071 ms22,769
—Nemotron 3.5 ASR StreamingTogether AI—1548 ms22,727
—Universal StreamingAssemblyAI—1513 ms20,215
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: 10th of 30 on Word Error Rate.

Latency vs accuracy

TTFS vs WEREach point is one measured model · Grok STT highlighted · 30-day averagesTime to Final Segment against Word Error Rate for every measured STT model, with Grok STT 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, 6.7%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, 3.2%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%Enhanced — 299 ms, 4.3%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 — 805 ms, 7.1%Chirp 2 — 873 ms, 4.9%Grok STT — 207 ms, 4.6%

Where the errors come from

WER compositionGrok STT's Word Error Rate split by error type · 30-day averageGrok STT's WER split into substitutions, deletions and insertions.
  • Grok STT4.6%
SubstitutionsDeletionsInsertions

Averages and tail latency

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

Grok STT latency distributionMilliseconds · shared axis across metrics · last 30 daysGrok STT's latency percentiles per metric: p25–p75 band, p50 tick, whisker to p99.
  • Time to Final Segmentp50 197 ms · p99 479 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 Segment207 ms167 ms197 ms221 ms248 ms269 ms479 ms22,833
Word Error Rate4.6%0.0%0.0%5.3%11.1%18.2%62.1%22,870

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 daysGrok STT's daily median Time to Final Segment over the last 30 days.
Grok STT · Sep 1: 186 msGrok STT · Sep 2: 200 msGrok STT · Sep 3: 193 msGrok STT · Sep 4: 190 msGrok STT · Sep 5: 184 msGrok STT · Sep 6: 186 msGrok STT · Sep 7: 186 msGrok STT · Sep 8: 202 msGrok STT · Sep 9: 214 msGrok STT · Sep 10: 218 msGrok STT · Sep 11: 200 msGrok STT · Sep 12: 198 msGrok STT · Sep 13: 193 msGrok STT · Sep 14: 217 msGrok STT · Sep 15: 192 ms
Word Error Rate — daily averageDaily average · UTC daysGrok STT's daily Word Error Rate over the last 30 days.
Grok STT · Sep 1: 5.5%Grok STT · Sep 2: 4.8%Grok STT · Sep 3: 4.5%Grok STT · Sep 4: 5.2%Grok STT · Sep 5: 4.4%Grok STT · Sep 6: 5.0%Grok STT · Sep 7: 9.1%Grok STT · Sep 8: 4.4%Grok STT · Sep 9: 8.2%Grok STT · Sep 10: 5.2%Grok STT · Sep 11: 4.5%Grok STT · Sep 12: 8.4%Grok STT · Sep 13: 4.8%Grok STT · Sep 14: 4.6%Grok STT · Sep 15: 3.9%

Time to Final Segment by dataset

Grok STT by test conditionMilliseconds · lower is better · best condition firstGrok STT's Time to Final Segment on each benchmark dataset.
  1. WildASR accents173 ms
  2. WildASR reverb186 ms
  3. WildASR phone codec197 ms
  4. WildASR clean204 ms
  5. WildASR noise gaps207 ms
  6. WildASR far-field208 ms
  7. PipeCat (production)210 ms
  8. WildASR clipping259 ms
Time to Final Segment per dataset, with the number of samples behind each figure.
DatasetTTFSSamples
LibriSpeech235 ms1
ProductionPipeCat210 ms11,497
AccentsWildASR173 ms1,115
CleanWildASR204 ms4,570
ClippingWildASR259 ms1,138
Far-fieldWildASR208 ms1,141
Noise gapsWildASR207 ms1,141
Phone codecWildASR197 ms1,136
ReverbWildASR186 ms1,094

Strongest condition: WildASR accents at 173 ms · weakest: WildASR clipping at 259 ms.

How fast is Grok STT?

On xAI, Grok STT measures mean 207 ms time to final segment (14th of 28). Last measured 2026-09-15.

How accurate is Grok STT?

On xAI, Grok STT measures 4.6% word error rate (10th of 30). Last measured 2026-09-15.

Who hosts Grok STT?

Grok STT is created by xAI and served by xAI. Coval measures each hosted endpoint separately.

Limits of this comparison

Coval measures both the text it returns and how quickly its streaming results arrive through xAI's official API.

  • The service is evolving; xAI's official documentation remains the source for current languages, limits and feature availability.

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