Voice AI BenchmarksBy

OPENAISPEECH-TO-TEXT106,500 SAMPLES / 30 DAYSLAST RUN OCT 14, 2026, 10:30 UTC

official resource

GPT-4o Transcribe speech-to-text benchmarks

GPT-4o Transcribe, hosted by OpenAI, measures mean 784 ms time to final segment (28th of 29) and 3.7% word error rate (9th of 31) among STT systems. Results cover the last 30 days. Coval benchmarks hosted endpoints daily using a consistent evaluation methodology. Last measured .

GPT-4o Transcribe is OpenAI's GPT-4o-based speech recognition model for the Audio API.

Time to Final Segment#28 / 29
784ms
Word Error Rate#9 / 31
3.7%

Overview

GPT-4o Transcribe is a dedicated transcription model rather than the transcription path inside an OpenAI Realtime session.

GPT-4o Transcribe 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
OpenAI
Hosted by
OpenAI
Source
Official API
Licensing
Proprietary
Deployment
Cloud
Region
US
Features
Keyterm biasing, Multilingual, VAD

How GPT-4o Transcribe 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 GPT-4o Transcribe highlighted.
  1. #1Qwen3 ASR 1.7b39 ms
  2. #2STT RT v545 ms
  3. #3Qwen3 ASR Fast47 ms
  4. #4STT 167 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
  13. #13Ink 2149 ms
  14. #14Universal 3.5 Pro168 ms
  15. #15Whisper Large v3via Together AI199 ms
  16. #16Pulse205 ms
  17. #17Grok STT223 ms
  18. #18resonant-1252 ms
  19. #19Default258 ms
  20. #20Linden 1271 ms
  21. #21Gemini 3.5 Transcribe Live317 ms
  22. #22Daytona STT399 ms
  23. #23Voxtral Mini Transcribe Realtime 2602411 ms
  24. #24GPT Realtime Whisper599 ms
  25. #25Solaria 1626 ms
  26. #26GPT-4o mini Transcribe708 ms
  27. #27Chirp 3768 ms
  28. #28GPT-4o Transcribe784 ms
Show all 29 models
  1. #29Chirp 2908 ms
median of all models · 199 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 AI67 ms1209 ms24,749
5Nova 3Deepgram85 ms1432 ms26,002
6Nova 2Deepgram87 ms1417 ms25,960
7Parakeet TDT 0.6B v3Together AI97 ms1228 ms26,461
8Flux MultilingualDeepgram98 ms1165 ms26,007
9FluxDeepgram99 ms1092 ms26,005
10Scribe v2 RealtimeElevenLabs122 ms2157 ms26,640
11Whisper Large v3Baseten131 ms924 ms13,663
12Universal 3.6 ProAssemblyAI135 ms1072 ms24,375
13Ink 2Cartesia149 ms1967 ms26,653
14Universal 3.5 ProAssemblyAI168 ms1026 ms26,654
15Whisper Large v3Together AI199 ms1261 ms26,446
16PulseSmallest205 ms1976 ms26,654
17Grok STTSpaceXAI223 ms—26,576
18resonant-1Reson8252 ms—26,653
19DefaultGradium258 ms1987 ms26,632
20Linden 1Speechmatics271 ms1537 ms26,595
21Gemini 3.5 Transcribe LiveGemini317 ms1607 ms26,566
22Daytona STTGuava399 ms5695 ms5,225
23Voxtral Mini Transcribe Realtime 2602Mistral411 ms1851 ms26,168
24GPT Realtime WhisperOpenAI599 ms1824 ms26,622
25Solaria 1Gladia626 ms1775 ms26,215
26GPT-4o mini TranscribeOpenAI708 ms—26,621
27Chirp 3Google768 ms5957 ms26,638
28GPT-4o TranscribeOpenAI784 ms—26,625
29Chirp 2Google908 ms6099 ms26,650
—Nemotron 3.5 ASR StreamingTogether AI—1544 ms26,549
—Universal StreamingAssemblyAI—1509 ms26,625
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 31 on Word Error Rate.

Latency vs accuracy

TTFS vs WEREach point is one measured model · GPT-4o Transcribe highlighted · 30-day averagesTime to Final Segment against Word Error Rate for every measured STT model, with GPT-4o Transcribe highlighted.
Qwen3 ASR 1.7b — 39 ms, 2.9%STT RT v5 — 45 ms, 4.8%Qwen3 ASR Fast — 47 ms, 2.6%STT 1 — 67 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 — 149 ms, 3.6%Universal 3.5 Pro — 168 ms, 2.3%Whisper Large v3 via Together AI — 199 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 — 411 ms, 5.3%GPT Realtime Whisper — 599 ms, 4.0%Solaria 1 — 626 ms, 6.0%GPT-4o mini Transcribe — 708 ms, 3.9%Chirp 3 — 768 ms, 3.2%Chirp 2 — 908 ms, 3.9%GPT-4o Transcribe — 784 ms, 3.7%

Where the errors come from

WER compositionGPT-4o Transcribe's Word Error Rate split by error type · 30-day averageGPT-4o Transcribe's WER split into substitutions, deletions and insertions.
  • GPT-4o Transcribe3.7%
SubstitutionsDeletionsInsertions

Averages and tail latency

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

GPT-4o Transcribe latency distributionMilliseconds · shared axis across metrics · last 30 daysGPT-4o Transcribe's latency percentiles per metric: p25–p75 band, p50 tick, whisker to max.
  • Time to Final Segmentp50 747 ms · Max 30943 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 Segment784 ms646 ms747 ms850 ms952 ms1041 ms30943 ms26,625
Word Error Rate3.7%0.0%0.0%4.8%12.5%20.0%100.0%26,625

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 daysGPT-4o Transcribe's daily median Time to Final Segment over the last 30 days.
GPT-4o Transcribe · Sep 14: 699 msGPT-4o Transcribe · Sep 15: 804 msGPT-4o Transcribe · Sep 16: 695 msGPT-4o Transcribe · Sep 17: 757 msGPT-4o Transcribe · Sep 18: 755 msGPT-4o Transcribe · Sep 19: 716 msGPT-4o Transcribe · Sep 20: 714 msGPT-4o Transcribe · Sep 21: 746 msGPT-4o Transcribe · Sep 22: 779 msGPT-4o Transcribe · Sep 23: 775 msGPT-4o Transcribe · Sep 24: 771 msGPT-4o Transcribe · Sep 25: 736 msGPT-4o Transcribe · Sep 26: 707 msGPT-4o Transcribe · Sep 27: 722 msGPT-4o Transcribe · Sep 28: 723 msGPT-4o Transcribe · Sep 29: 780 msGPT-4o Transcribe · Sep 30: 699 msGPT-4o Transcribe · Oct 1: 802 msGPT-4o Transcribe · Oct 2: 789 msGPT-4o Transcribe · Oct 3: 717 msGPT-4o Transcribe · Oct 4: 643 msGPT-4o Transcribe · Oct 5: 744 msGPT-4o Transcribe · Oct 6: 824 msGPT-4o Transcribe · Oct 7: 769 msGPT-4o Transcribe · Oct 8: 695 msGPT-4o Transcribe · Oct 9: 721 msGPT-4o Transcribe · Oct 10: 683 msGPT-4o Transcribe · Oct 11: 769 msGPT-4o Transcribe · Oct 12: 729 msGPT-4o Transcribe · Oct 13: 759 msGPT-4o Transcribe · Oct 14: 703 ms
Word Error Rate — daily averageDaily average · UTC daysGPT-4o Transcribe's daily Word Error Rate over the last 30 days.
GPT-4o Transcribe · Sep 14: 4.7%GPT-4o Transcribe · Sep 15: 5.5%GPT-4o Transcribe · Sep 16: 3.8%GPT-4o Transcribe · Sep 17: 5.2%GPT-4o Transcribe · Sep 18: 4.7%GPT-4o Transcribe · Sep 19: 4.6%GPT-4o Transcribe · Sep 20: 6.5%GPT-4o Transcribe · Sep 21: 4.0%GPT-4o Transcribe · Sep 22: 4.6%GPT-4o Transcribe · Sep 23: 5.0%GPT-4o Transcribe · Sep 24: 5.0%GPT-4o Transcribe · Sep 25: 6.5%GPT-4o Transcribe · Sep 26: 5.6%GPT-4o Transcribe · Sep 27: 5.6%GPT-4o Transcribe · Sep 28: 3.4%GPT-4o Transcribe · Sep 29: 4.6%GPT-4o Transcribe · Sep 30: 4.6%GPT-4o Transcribe · Oct 1: 4.8%GPT-4o Transcribe · Oct 2: 4.2%GPT-4o Transcribe · Oct 3: 3.6%GPT-4o Transcribe · Oct 4: 6.0%GPT-4o Transcribe · Oct 5: 3.8%GPT-4o Transcribe · Oct 6: 5.5%GPT-4o Transcribe · Oct 7: 5.0%GPT-4o Transcribe · Oct 8: 5.0%GPT-4o Transcribe · Oct 9: 5.0%GPT-4o Transcribe · Oct 10: 5.1%GPT-4o Transcribe · Oct 11: 4.3%GPT-4o Transcribe · Oct 12: 4.6%GPT-4o Transcribe · Oct 13: 4.7%GPT-4o Transcribe · Oct 14: 3.9%

Time to Final Segment by dataset

GPT-4o Transcribe by test conditionMilliseconds · lower is better · best condition firstGPT-4o Transcribe's Time to Final Segment on each benchmark dataset.
  1. WildASR accents673 ms
  2. WildASR clean784 ms
  3. PipeCat (production)790 ms
  4. WildASR far-field791 ms
  5. WildASR phone codec793 ms
  6. WildASR noise gaps795 ms
  7. WildASR clipping796 ms
Time to Final Segment per dataset, with the number of samples behind each figure.
DatasetTTFSSamples
ProductionPipeCat790 ms14,006
AccentsWildASR673 ms1,402
CleanWildASR784 ms5,607
ClippingWildASR796 ms1,402
Far-fieldWildASR791 ms1,404
Noise gapsWildASR795 ms1,402
Phone codecWildASR793 ms1,402

Strongest condition: WildASR accents at 673 ms · weakest: WildASR clipping at 796 ms.

How fast is GPT-4o Transcribe?

On OpenAI, GPT-4o Transcribe measures mean 784 ms time to final segment (28th of 29). Last measured 2026-10-14.

How accurate is GPT-4o Transcribe?

On OpenAI, GPT-4o Transcribe measures 3.7% word error rate (9th of 31). Last measured 2026-10-14.

Who hosts GPT-4o Transcribe?

GPT-4o Transcribe is created by OpenAI and served by OpenAI. Coval measures each hosted endpoint separately.

Limits of this comparison

Coval reports accuracy, first partial output and final-segment timing separately because a fast partial does not guarantee a fast settled transcript.

  • Prompting, language mix and audio format can affect production behavior beyond Coval's fixed public datasets.

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