Voice AI BenchmarksBy

OPENAISPEECH-TO-TEXT200,489 SAMPLES / 30 DAYSLAST RUN OCT 14, 2026, 10:30 UTC

official resource

Whisper Large v3 speech-to-text benchmarks

Whisper Large v3, hosted by Baseten (dedicated inference), measures mean 131 ms time to final segment (11th of 29) and 4.5% word error rate (16th of 31) among STT systems. Results cover the last 30 days. Coval benchmarks hosted endpoints daily using a consistent evaluation methodology. Last measured .

Whisper Large v3 is OpenAI's downloadable multilingual recognizer.

Time to Final Segment#11 / 29
131ms
Word Error Rate#16 / 31
4.5%
Time to First Token#1 / 27
924ms

Overview

OpenAI publishes Whisper's code and model weights, while Together AI and Baseten each supply a measured inference endpoint.

Whisper Large v3 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
Baseten, Together AI
Source
Dedicated inference
Licensing
Open-weight
Deployment
On-prem
Region
US
Features
Multilingual, VAD

How Whisper Large v3 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 Whisper Large v3 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
Show all 29 models
  1. #16Pulse205 ms
  2. #17Grok STT223 ms
  3. #18resonant-1252 ms
  4. #19Default258 ms
  5. #20Linden 1271 ms
  6. #21Gemini 3.5 Transcribe Live317 ms
  7. #22Daytona STT399 ms
  8. #23Voxtral Mini Transcribe Realtime 2602411 ms
  9. #24GPT Realtime Whisper599 ms
  10. #25Solaria 1626 ms
  11. #26GPT-4o mini Transcribe708 ms
  12. #27Chirp 3768 ms
  13. #28GPT-4o Transcribe784 ms
  14. #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.

Whisper Large v3 runs on dedicated inference, ranked here against shared endpoints. Highest relative placement: 1st of 27 on Time to First Token.

Latency vs accuracy

TTFS vs WEREach point is one measured model · Whisper Large v3 highlighted · 30-day averagesTime to Final Segment against Word Error Rate for every measured STT model, with Whisper Large v3 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%Universal 3.6 Pro — 135 ms, 2.1%Ink 2 — 149 ms, 3.6%Universal 3.5 Pro — 168 ms, 2.3%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%GPT-4o Transcribe — 784 ms, 3.7%Chirp 2 — 908 ms, 3.9%Whisper Large v3 via Baseten — 131 ms, 4.5%Whisper Large v3 via Together AI — 199 ms, 7.4%

Where the errors come from

WER compositionWhisper Large v3's Word Error Rate split by error type · 30-day averageWhisper Large v3's WER split into substitutions, deletions and insertions.
  • Whisper Large v3via Baseten4.5%
  • Whisper Large v3via Together AI7.4%
SubstitutionsDeletionsInsertions

Averages and tail latency

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

Whisper Large v3 latency distributionMilliseconds · shared axis across metrics · last 30 daysWhisper Large v3's latency percentiles per metric: p25–p75 band, p50 tick, whisker to max.
  • Time to Final Segmentvia Basetenp50 114 ms · Max 10654 ms
  • Time to Final Segmentvia Together AIp50 111 ms · Max 12594 ms
  • Time to First Tokenvia Basetenp50 1003 ms · Max 4862 ms
  • Time to First Tokenvia Together AIp50 1301 ms · Max 17738 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.
MetricHostAveragep25p50p75p90p95MaxSamples
Time to Final SegmentBaseten131 ms83 ms114 ms153 ms204 ms252 ms10654 ms13,663
Word Error RateBaseten4.5%0.0%0.0%7.7%13.3%20.0%100.0%13,663
Time to First TokenBaseten924 ms609 ms1003 ms1072 ms1327 ms1613 ms4862 ms13,663
Time to Final SegmentTogether AI199 ms84 ms111 ms174 ms311 ms686 ms12594 ms26,446
Word Error RateTogether AI7.4%0.0%4.8%11.1%18.8%26.3%126.7%26,446
Time to First TokenTogether AI1261 ms859 ms1301 ms1459 ms1865 ms2311 ms17738 ms26,390

Last 30 days

Daily medians from the same measurement runs · gaps are days without qualifying runs.

Time to Final Segment — daily p50Line p50 per host · UTC daysWhisper Large v3's daily median Time to Final Segment over the last 30 days.
Whisper Large v3 · Sep 15: 119 msWhisper Large v3 · Sep 16: 111 msWhisper Large v3 · Sep 17: 110 msWhisper Large v3 · Sep 18: 124 msWhisper Large v3 · Sep 19: 106 msWhisper Large v3 · Sep 20: 109 msWhisper Large v3 · Sep 21: 106 msWhisper Large v3 · Sep 22: 111 msWhisper Large v3 · Sep 23: 139 msWhisper Large v3 · Sep 24: 120 msWhisper Large v3 · Sep 25: 122 msWhisper Large v3 · Sep 26: 128 msWhisper Large v3 · Sep 27: 121 msWhisper Large v3 · Sep 28: 107 msWhisper Large v3 · Sep 29: 123 msWhisper Large v3 · Sep 30: 107 msWhisper Large v3 · Oct 1: 117 msWhisper Large v3 · Oct 2: 110 msWhisper Large v3 · Oct 3: 110 msWhisper Large v3 · Oct 4: 118 msWhisper Large v3 · Oct 5: 112 msWhisper Large v3 · Oct 6: 111 msWhisper Large v3 · Oct 7: 108 msWhisper Large v3 · Oct 8: 111 msWhisper Large v3 · Oct 9: 111 msWhisper Large v3 · Oct 10: 112 msWhisper Large v3 · Oct 11: 105 msWhisper Large v3 · Oct 12: 108 msWhisper Large v3 · Oct 13: 112 msWhisper Large v3 · Oct 14: 126 msWhisper Large v3 · Sep 14: 141 msWhisper Large v3 · Sep 15: 107 msWhisper Large v3 · Sep 16: 140 msWhisper Large v3 · Sep 17: 130 msWhisper Large v3 · Sep 18: 108 msWhisper Large v3 · Sep 19: 95 msWhisper Large v3 · Sep 20: 111 msWhisper Large v3 · Sep 21: 131 msWhisper Large v3 · Sep 22: 122 msWhisper Large v3 · Sep 23: 122 msWhisper Large v3 · Sep 24: 111 msWhisper Large v3 · Sep 25: 130 msWhisper Large v3 · Sep 26: 102 msWhisper Large v3 · Sep 27: 104 msWhisper Large v3 · Sep 28: 145 msWhisper Large v3 · Sep 29: 121 msWhisper Large v3 · Sep 30: 170 msWhisper Large v3 · Oct 1: 91 msWhisper Large v3 · Oct 2: 111 msWhisper Large v3 · Oct 3: 121 msWhisper Large v3 · Oct 4: 101 msWhisper Large v3 · Oct 5: 107 msWhisper Large v3 · Oct 6: 128 msWhisper Large v3 · Oct 7: 112 msWhisper Large v3 · Oct 8: 119 msWhisper Large v3 · Oct 9: 136 msWhisper Large v3 · Oct 10: 153 msWhisper Large v3 · Oct 11: 130 msWhisper Large v3 · Oct 12: 115 msWhisper Large v3 · Oct 13: 104 msWhisper Large v3 · Oct 14: 106 ms
Whisper Large v3via BasetenWhisper Large v3via Together AI
Word Error Rate — daily averageDaily average · UTC daysWhisper Large v3's daily Word Error Rate over the last 30 days.
Whisper Large v3 · Sep 15: 5.3%Whisper Large v3 · Sep 16: 5.1%Whisper Large v3 · Sep 17: 4.8%Whisper Large v3 · Sep 18: 5.1%Whisper Large v3 · Sep 19: 4.9%Whisper Large v3 · Sep 20: 4.2%Whisper Large v3 · Sep 21: 5.0%Whisper Large v3 · Sep 22: 4.4%Whisper Large v3 · Sep 23: 4.8%Whisper Large v3 · Sep 24: 5.1%Whisper Large v3 · Sep 25: 5.1%Whisper Large v3 · Sep 26: 4.8%Whisper Large v3 · Sep 27: 4.6%Whisper Large v3 · Sep 28: 5.1%Whisper Large v3 · Sep 29: 4.6%Whisper Large v3 · Sep 30: 4.4%Whisper Large v3 · Oct 1: 4.2%Whisper Large v3 · Oct 2: 5.2%Whisper Large v3 · Oct 3: 4.9%Whisper Large v3 · Oct 4: 4.5%Whisper Large v3 · Oct 5: 4.9%Whisper Large v3 · Oct 6: 4.8%Whisper Large v3 · Oct 7: 5.0%Whisper Large v3 · Oct 8: 5.0%Whisper Large v3 · Oct 9: 4.9%Whisper Large v3 · Oct 10: 5.5%Whisper Large v3 · Oct 11: 5.0%Whisper Large v3 · Oct 12: 5.6%Whisper Large v3 · Oct 13: 5.5%Whisper Large v3 · Oct 14: 5.0%Whisper Large v3 · Sep 14: 9.7%Whisper Large v3 · Sep 15: 9.0%Whisper Large v3 · Sep 16: 9.4%Whisper Large v3 · Sep 17: 9.1%Whisper Large v3 · Sep 18: 8.4%Whisper Large v3 · Sep 19: 8.1%Whisper Large v3 · Sep 20: 8.6%Whisper Large v3 · Sep 21: 9.5%Whisper Large v3 · Sep 22: 8.9%Whisper Large v3 · Sep 23: 8.2%Whisper Large v3 · Sep 24: 9.3%Whisper Large v3 · Sep 25: 8.6%Whisper Large v3 · Sep 26: 7.5%Whisper Large v3 · Sep 27: 8.2%Whisper Large v3 · Sep 28: 8.5%Whisper Large v3 · Sep 29: 8.5%Whisper Large v3 · Sep 30: 8.7%Whisper Large v3 · Oct 1: 7.6%Whisper Large v3 · Oct 2: 7.7%Whisper Large v3 · Oct 3: 8.7%Whisper Large v3 · Oct 4: 8.1%Whisper Large v3 · Oct 5: 8.8%Whisper Large v3 · Oct 6: 9.6%Whisper Large v3 · Oct 7: 7.3%Whisper Large v3 · Oct 8: 8.7%Whisper Large v3 · Oct 9: 7.4%Whisper Large v3 · Oct 10: 7.0%Whisper Large v3 · Oct 11: 8.4%Whisper Large v3 · Oct 12: 7.6%Whisper Large v3 · Oct 13: 8.2%Whisper Large v3 · Oct 14: 7.5%
Whisper Large v3via BasetenWhisper Large v3via Together AI

Time to Final Segment by dataset

Whisper Large v3 by test conditionMilliseconds · lower is better · best condition firstWhisper Large v3's Time to Final Segment on each benchmark dataset.
  1. WildASR accents90 ms
  2. WildASR noise gaps116 ms
  3. WildASR clean128 ms
  4. WildASR phone codec132 ms
  5. PipeCat (production)137 ms
  6. WildASR far-field137 ms
  7. WildASR clipping137 ms
Time to Final Segment per dataset, with the number of samples behind each figure.
DatasetHostTTFSSamples
ProductionPipeCatBaseten137 ms7,192
AccentsWildASRBaseten90 ms720
CleanWildASRBaseten128 ms2,871
ClippingWildASRBaseten137 ms720
Far-fieldWildASRBaseten137 ms720
Noise gapsWildASRBaseten116 ms720
Phone codecWildASRBaseten132 ms720
ProductionPipeCatTogether AI201 ms13,908
AccentsWildASRTogether AI216 ms1,391
CleanWildASRTogether AI196 ms5,576
ClippingWildASRTogether AI209 ms1,390
Far-fieldWildASRTogether AI193 ms1,397
Noise gapsWildASRTogether AI173 ms1,391
Phone codecWildASRTogether AI197 ms1,393

Strongest condition: WildASR accents at 90 ms · weakest: WildASR clipping at 137 ms.

How fast is Whisper Large v3?

On Baseten (dedicated inference), Whisper Large v3 measures mean 131 ms time to final segment (11th of 29) and mean 924 ms time to first token (1st of 27). Last measured 2026-10-14. On Together AI, Whisper Large v3 measures mean 199 ms time to final segment (15th of 29) and mean 1261 ms time to first token (9th of 27). Last measured 2026-10-14.

How accurate is Whisper Large v3?

On Baseten (dedicated inference), Whisper Large v3 measures 4.5% word error rate (16th of 31). Last measured 2026-10-14. On Together AI, Whisper Large v3 measures 7.4% word error rate (27th of 31). Last measured 2026-10-14.

Who hosts Whisper Large v3?

Whisper Large v3 is created by OpenAI and served by Baseten and Together AI. Coval measures each hosted endpoint separately.

Limits of this comparison

Coval measures it on two hosts, Together AI's shared API and a Baseten dedicated streaming endpoint, so each latency figure applies to that host's serving configuration.

  • Quantization, hardware, batching and decoding parameters can change speed and accuracy for an open-weight model. Baseten's endpoint is dedicated inference, so its latency is not ranked against shared APIs here.

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