Voice AI BenchmarksBy

ASSEMBLYAISPEECH-TO-TEXT100,990 SAMPLES / 30 DAYSLAST RUN SEP 15, 2026, 07:00 UTC

official resource

Universal 3.5 Pro speech-to-text benchmarks

Universal 3.5 Pro, hosted by AssemblyAI (Assembly AI, assemblyai.com), measures mean 173 ms time to final segment (13th of 28) and 3.1% word error rate (1st of 30) among STT systems. Results cover the last 30 days. Last measured .

Universal 3.5 Pro is AssemblyAI's accuracy-oriented Universal recognizer.

Time to Final Segment#13 / 28
173ms
Word Error Rate#1 / 30
3.1%
Time to First Token#3 / 26
1040ms

Overview

The tested AssemblyAI endpoint supports multilingual recognition, diarization, code switching and keyterm controls.

Universal 3.5 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 3.5 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 3.5 Pro 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
Show all 28 models
  1. #14Grok STT207 ms
  2. #15Defaultvia Speechmatics209 ms
  3. #16Pulse212 ms
  4. #17Defaultvia Gradium246 ms
  5. #18resonant-1264 ms
  6. #19Whisper Large v3via Together AI296 ms
  7. #20Enhanced299 ms
  8. #21Gemini 3.5 Transcribe Live305 ms
  9. #22Voxtral Mini Transcribe Realtime 2602400 ms
  10. #23GPT Realtime Whisper551 ms
  11. #24GPT-4o mini Transcribe690 ms
  12. #25GPT-4o Transcribe754 ms
  13. #26Chirp 3776 ms
  14. #27Solaria 1802 ms
  15. #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 ms4,371
3STT RT v5Soniox57 ms1529 ms22,468
4STT 1Inworld AI65 ms1400 ms22,756
5Parakeet TDT 0.6B v3Together AI81 ms1215 ms22,408
6Nova 3Deepgram89 ms1418 ms22,775
7Nova 2Deepgram92 ms1419 ms22,653
8Flux MultilingualDeepgram98 ms1163 ms12,171
9FluxDeepgram99 ms1076 ms12,163
10Ink 2Cartesia122 ms1827 ms22,782
11Whisper Large v3Baseten125 ms912 ms1,252
12Scribe v2 RealtimeElevenLabs133 ms2175 ms22,786
13Universal 3.5 ProAssemblyAI173 ms1040 ms20,173
14Grok STTxAI207 ms—22,773
15DefaultSpeechmatics209 ms1428 ms22,796
16PulseSmallest212 ms2011 ms22,776
17DefaultGradium246 ms1980 ms22,759
18resonant-1Reson8264 ms—22,780
19Whisper Large v3Together AI296 ms1338 ms22,665
20EnhancedSpeechmatics299 ms1492 ms22,796
21Gemini 3.5 Transcribe LiveGemini305 ms1659 ms16,323
22Voxtral Mini Transcribe Realtime 2602Mistral400 ms1850 ms22,506
23GPT Realtime WhisperOpenAI551 ms1814 ms22,732
24GPT-4o mini TranscribeOpenAI690 ms—22,750
25GPT-4o TranscribeOpenAI754 ms—22,751
26Chirp 3Google776 ms5974 ms22,795
27Solaria 1Gladia802 ms1800 ms22,374
28Chirp 2Google873 ms6072 ms22,709
—Nemotron 3.5 ASR StreamingTogether AI—1547 ms22,667
—Universal StreamingAssemblyAI—1513 ms20,155
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 30 on Word Error Rate.

Latency vs accuracy

TTFS vs WEREach point is one measured model · Universal 3.5 Pro highlighted · 30-day averagesTime to Final Segment against Word Error Rate for every measured STT model, with Universal 3.5 Pro 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, 6.6%Ink 2 — 122 ms, 4.8%Whisper Large v3 via Baseten — 125 ms, 5.6%Scribe v2 Realtime — 133 ms, 5.5%Grok STT — 207 ms, 4.6%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 — 802 ms, 7.1%Chirp 2 — 873 ms, 4.9%Universal 3.5 Pro — 173 ms, 3.1%

Where the errors come from

WER compositionUniversal 3.5 Pro's Word Error Rate split by error type · 30-day averageUniversal 3.5 Pro's WER split into substitutions, deletions and insertions.
  • Universal 3.5 Pro3.1%
SubstitutionsDeletionsInsertions

Averages and tail latency

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

Universal 3.5 Pro latency distributionMilliseconds · shared axis across metrics · last 30 daysUniversal 3.5 Pro's latency percentiles per metric: p25–p75 band, p50 tick, whisker to p99.
  • Time to Final Segmentp50 143 ms · p99 581 ms
  • Time to First Tokenp50 1117 ms · p99 2938 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 Segment173 ms105 ms143 ms196 ms278 ms357 ms581 ms20,173
Word Error Rate3.1%0.0%0.0%3.4%9.1%13.3%36.3%20,204
Time to First Token1040 ms717 ms1117 ms1218 ms1520 ms1820 ms2938 ms20,205

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 3.5 Pro's daily median Time to Final Segment over the last 30 days.
Universal 3.5 Pro · Sep 1: 174 msUniversal 3.5 Pro · Sep 2: 141 msUniversal 3.5 Pro · Sep 3: 142 msUniversal 3.5 Pro · Sep 4: 144 msUniversal 3.5 Pro · Sep 5: 141 msUniversal 3.5 Pro · Sep 8: 207 msUniversal 3.5 Pro · Sep 9: 161 msUniversal 3.5 Pro · Sep 10: 152 msUniversal 3.5 Pro · Sep 11: 151 msUniversal 3.5 Pro · Sep 12: 124 msUniversal 3.5 Pro · Sep 13: 141 msUniversal 3.5 Pro · Sep 14: 151 msUniversal 3.5 Pro · Sep 15: 133 ms
Word Error Rate — daily averageDaily average · UTC daysUniversal 3.5 Pro's daily Word Error Rate over the last 30 days.
Universal 3.5 Pro · Sep 1: 3.4%Universal 3.5 Pro · Sep 2: 3.7%Universal 3.5 Pro · Sep 3: 3.1%Universal 3.5 Pro · Sep 4: 3.3%Universal 3.5 Pro · Sep 5: 3.4%Universal 3.5 Pro · Sep 8: 2.8%Universal 3.5 Pro · Sep 9: 3.6%Universal 3.5 Pro · Sep 10: 3.6%Universal 3.5 Pro · Sep 11: 3.0%Universal 3.5 Pro · Sep 12: 3.6%Universal 3.5 Pro · Sep 13: 3.7%Universal 3.5 Pro · Sep 14: 3.9%Universal 3.5 Pro · Sep 15: 3.6%

Time to Final Segment by dataset

Universal 3.5 Pro by test conditionMilliseconds · lower is better · best condition firstUniversal 3.5 Pro's Time to Final Segment on each benchmark dataset.
  1. WildASR clean166 ms
  2. WildASR phone codec169 ms
  3. WildASR reverb170 ms
  4. WildASR noise gaps171 ms
  5. PipeCat (production)173 ms
  6. WildASR clipping176 ms
  7. WildASR far-field185 ms
  8. WildASR accents201 ms
Time to Final Segment per dataset, with the number of samples behind each figure.
DatasetTTFSSamples
LibriSpeech222 ms1
ProductionPipeCat173 ms10,166
AccentsWildASR201 ms982
CleanWildASR166 ms4,037
ClippingWildASR176 ms1,005
Far-fieldWildASR185 ms1,008
Noise gapsWildASR171 ms1,007
Phone codecWildASR169 ms1,003
ReverbWildASR170 ms964

Strongest condition: WildASR clean at 166 ms · weakest: LibriSpeech at 222 ms.

How fast is Universal 3.5 Pro?

On AssemblyAI, Universal 3.5 Pro measures mean 173 ms time to final segment (13th of 28) and mean 1040 ms time to first token (3rd of 26). Last measured 2026-09-15.

How accurate is Universal 3.5 Pro?

On AssemblyAI, Universal 3.5 Pro measures 3.1% word error rate (1st of 30). Last measured 2026-09-15.

Who hosts Universal 3.5 Pro?

Universal 3.5 Pro is created by AssemblyAI and served by AssemblyAI. Coval measures each hosted endpoint separately.

Limits of this comparison

Coval keeps it independent from Universal Streaming so different product modes are not collapsed into one brand result.

  • Timing reflects Coval's API mode and should not be assumed to describe every asynchronous transcription workflow.

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