Voice AI BenchmarksBy

ASSEMBLYAISPEECH-TO-TEXT120,545 SAMPLES / 30 DAYSLAST RUN OCT 14, 2026, 03:30 UTC

Universal 3.6 Pro speech-to-text benchmarks

Universal 3.6 Pro, hosted by AssemblyAI (Assembly AI, assemblyai.com), measures mean 135 ms time to final segment (12th of 29) and 2.1% word error rate (1st of 31) among STT systems. Results cover the last 30 days. Last measured .

Coval has active speech-to-text measurements for Universal 3.6 Pro, created by AssemblyAI and served through AssemblyAI's own API.

Time to Final Segment#12 / 29
135ms
Word Error Rate#1 / 31
2.1%
Time to First Token#4 / 27
1072ms

Overview

Universal 3.6 Pro 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
AssemblyAI
Hosted by
AssemblyAI
Source
Official API
Licensing
Proprietary
Deployment
On-prem
Region
US
Features
Code switching, Diarization, Keyterm biasing, Multilingual, VAD

How Universal 3.6 Pro 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 Universal 3.6 Pro highlighted.
  1. #1Qwen3 ASR 1.7b39 ms
  2. #2STT RT v545 ms
  3. #3Qwen3 ASR Fast47 ms
  4. #4STT 166 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
Show all 29 models
  1. #13Ink 2148 ms
  2. #14Universal 3.5 Pro167 ms
  3. #15Whisper Large v3via Together AI200 ms
  4. #16Pulse205 ms
  5. #17Grok STT223 ms
  6. #18resonant-1252 ms
  7. #19Default257 ms
  8. #20Linden 1270 ms
  9. #21Gemini 3.5 Transcribe Live317 ms
  10. #22Daytona STT399 ms
  11. #23Voxtral Mini Transcribe Realtime 2602410 ms
  12. #24GPT Realtime Whisper599 ms
  13. #25Solaria 1639 ms
  14. #26GPT-4o mini Transcribe708 ms
  15. #27Chirp 3768 ms
  16. #28GPT-4o Transcribe783 ms
  17. #29Chirp 2908 ms
median of all models · 200 ms
STT models over the last 30 days, ranked on Time to Final Segment.
#ModelHostTTFSWERTTFTSamples
1Qwen3 ASR 1.7bBaseten39 ms945 ms13,235
2STT RT v5Soniox45 ms1518 ms23,533
3Qwen3 ASR FastNari Labs47 ms1752 ms26,652
4STT 1Inworld AI66 ms1213 ms24,749
5Nova 3Deepgram85 ms1431 ms26,002
6Nova 2Deepgram87 ms1417 ms25,960
7Parakeet TDT 0.6B v3Together AI97 ms1229 ms26,460
8Flux MultilingualDeepgram98 ms1165 ms26,007
9FluxDeepgram99 ms1091 ms26,005
10Scribe v2 RealtimeElevenLabs122 ms2158 ms26,640
11Whisper Large v3Baseten131 ms923 ms13,229
12Universal 3.6 ProAssemblyAI135 ms1072 ms24,109
13Ink 2Cartesia148 ms1965 ms26,652
14Universal 3.5 ProAssemblyAI167 ms1025 ms26,654
15Whisper Large v3Together AI200 ms1262 ms26,445
16PulseSmallest205 ms1976 ms26,654
17Grok STTSpaceXAI223 ms—26,576
18resonant-1Reson8252 ms—26,653
19DefaultGradium257 ms1986 ms26,632
20Linden 1Speechmatics270 ms1537 ms26,595
21Gemini 3.5 Transcribe LiveGemini317 ms1609 ms26,567
22Daytona STTGuava399 ms5700 ms4,794
23Voxtral Mini Transcribe Realtime 2602Mistral410 ms1852 ms26,199
24GPT Realtime WhisperOpenAI599 ms1824 ms26,622
25Solaria 1Gladia639 ms1785 ms26,195
26GPT-4o mini TranscribeOpenAI708 ms—26,621
27Chirp 3Google768 ms5956 ms26,638
28GPT-4o TranscribeOpenAI783 ms—26,625
29Chirp 2Google908 ms6098 ms26,650
—Nemotron 3.5 ASR StreamingTogether AI—1544 ms26,542
—Universal StreamingAssemblyAI—1508 ms26,623
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: 1st of 31 on Word Error Rate.

Latency vs accuracy

TTFS vs WEREach point is one measured model · Universal 3.6 Pro highlighted · 30-day averagesTime to Final Segment against Word Error Rate for every measured STT model, with Universal 3.6 Pro 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 — 66 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%Ink 2 — 148 ms, 3.6%Universal 3.5 Pro — 167 ms, 2.4%Whisper Large v3 via Together AI — 200 ms, 7.4%Pulse — 205 ms, 4.6%Grok STT — 223 ms, 4.1%resonant-1 — 252 ms, 2.6%Default — 257 ms, 8.8%Linden 1 — 270 ms, 3.9%Gemini 3.5 Transcribe Live — 317 ms, 3.0%Daytona STT — 399 ms, 5.0%Voxtral Mini Transcribe Realtime 2602 — 410 ms, 5.4%GPT Realtime Whisper — 599 ms, 4.0%Solaria 1 — 639 ms, 6.1%GPT-4o mini Transcribe — 708 ms, 3.9%Chirp 3 — 768 ms, 3.2%GPT-4o Transcribe — 783 ms, 3.8%Chirp 2 — 908 ms, 3.9%Universal 3.6 Pro — 135 ms, 2.1%

Where the errors come from

WER compositionUniversal 3.6 Pro's Word Error Rate split by error type · 30-day averageUniversal 3.6 Pro's WER split into substitutions, deletions and insertions.
  • Universal 3.6 Pro2.1%
SubstitutionsDeletionsInsertions

Averages and tail latency

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

Universal 3.6 Pro latency distributionMilliseconds · shared axis across metrics · last 30 daysUniversal 3.6 Pro's latency percentiles per metric: p25–p75 band, p50 tick, whisker to max.
  • Time to Final Segmentp50 124 ms · Max 13106 ms
  • Time to First Tokenp50 1118 ms · Max 5521 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 Segment135 ms93 ms124 ms161 ms204 ms235 ms13106 ms24,109
Word Error Rate2.1%0.0%0.0%3.2%8.3%12.5%60.0%24,109
Time to First Token1072 ms717 ms1118 ms1246 ms1618 ms1918 ms5521 ms24,109

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 daysUniversal 3.6 Pro's daily median Time to Final Segment over the last 30 days.
Universal 3.6 Pro · Sep 16: 159 msUniversal 3.6 Pro · Sep 17: 106 msUniversal 3.6 Pro · Sep 18: 128 msUniversal 3.6 Pro · Sep 19: 124 msUniversal 3.6 Pro · Sep 20: 126 msUniversal 3.6 Pro · Sep 21: 107 msUniversal 3.6 Pro · Sep 22: 143 msUniversal 3.6 Pro · Sep 23: 123 msUniversal 3.6 Pro · Sep 24: 123 msUniversal 3.6 Pro · Sep 25: 125 msUniversal 3.6 Pro · Sep 26: 125 msUniversal 3.6 Pro · Sep 27: 108 msUniversal 3.6 Pro · Sep 28: 143 msUniversal 3.6 Pro · Sep 29: 136 msUniversal 3.6 Pro · Sep 30: 120 msUniversal 3.6 Pro · Oct 1: 155 msUniversal 3.6 Pro · Oct 2: 130 msUniversal 3.6 Pro · Oct 3: 155 msUniversal 3.6 Pro · Oct 4: 157 msUniversal 3.6 Pro · Oct 5: 135 msUniversal 3.6 Pro · Oct 6: 173 msUniversal 3.6 Pro · Oct 7: 148 msUniversal 3.6 Pro · Oct 8: 144 msUniversal 3.6 Pro · Oct 9: 94 msUniversal 3.6 Pro · Oct 10: 100 msUniversal 3.6 Pro · Oct 11: 95 msUniversal 3.6 Pro · Oct 12: 97 msUniversal 3.6 Pro · Oct 13: 101 msUniversal 3.6 Pro · Oct 14: 91 ms
Word Error Rate — daily averageDaily average · UTC daysUniversal 3.6 Pro's daily Word Error Rate over the last 30 days.
Universal 3.6 Pro · Sep 16: 1.0%Universal 3.6 Pro · Sep 17: 3.1%Universal 3.6 Pro · Sep 18: 2.6%Universal 3.6 Pro · Sep 19: 2.1%Universal 3.6 Pro · Sep 20: 3.7%Universal 3.6 Pro · Sep 21: 3.0%Universal 3.6 Pro · Sep 22: 3.9%Universal 3.6 Pro · Sep 23: 3.2%Universal 3.6 Pro · Sep 24: 3.1%Universal 3.6 Pro · Sep 25: 3.4%Universal 3.6 Pro · Sep 26: 2.6%Universal 3.6 Pro · Sep 27: 3.2%Universal 3.6 Pro · Sep 28: 2.7%Universal 3.6 Pro · Sep 29: 2.4%Universal 3.6 Pro · Sep 30: 2.6%Universal 3.6 Pro · Oct 1: 2.3%Universal 3.6 Pro · Oct 2: 1.8%Universal 3.6 Pro · Oct 3: 3.2%Universal 3.6 Pro · Oct 4: 3.1%Universal 3.6 Pro · Oct 5: 2.4%Universal 3.6 Pro · Oct 6: 3.2%Universal 3.6 Pro · Oct 7: 2.9%Universal 3.6 Pro · Oct 8: 3.0%Universal 3.6 Pro · Oct 9: 2.6%Universal 3.6 Pro · Oct 10: 3.0%Universal 3.6 Pro · Oct 11: 2.9%Universal 3.6 Pro · Oct 12: 3.0%Universal 3.6 Pro · Oct 13: 3.2%Universal 3.6 Pro · Oct 14: 2.8%

Time to Final Segment by dataset

Universal 3.6 Pro by test conditionMilliseconds · lower is better · best condition firstUniversal 3.6 Pro's Time to Final Segment on each benchmark dataset.
  1. WildASR accents124 ms
  2. WildASR noise gaps133 ms
  3. WildASR clean134 ms
  4. WildASR phone codec136 ms
  5. PipeCat (production)137 ms
  6. WildASR far-field138 ms
  7. WildASR clipping139 ms
Time to Final Segment per dataset, with the number of samples behind each figure.
DatasetTTFSSamples
ProductionPipeCat137 ms12,667
AccentsWildASR124 ms1,269
CleanWildASR134 ms5,070
ClippingWildASR139 ms1,272
Far-fieldWildASR138 ms1,272
Noise gapsWildASR133 ms1,268
Phone codecWildASR136 ms1,272

Strongest condition: WildASR accents at 124 ms · weakest: WildASR clipping at 139 ms.

How fast is Universal 3.6 Pro?

On AssemblyAI, Universal 3.6 Pro measures mean 135 ms time to final segment (12th of 29) and mean 1072 ms time to first token (4th of 27). Last measured 2026-10-14.

How accurate is Universal 3.6 Pro?

On AssemblyAI, Universal 3.6 Pro measures 2.1% word error rate (1st of 31). Last measured 2026-10-14.

Who hosts Universal 3.6 Pro?

Universal 3.6 Pro is created by AssemblyAI and served by AssemblyAI. Coval measures each hosted endpoint separately.

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