Voice AI BenchmarksBy

DEEPGRAMSPEECH-TO-TEXT130,025 SAMPLES / 30 DAYSLAST RUN OCT 14, 2026, 06:00 UTC

official resource

Flux speech-to-text benchmarks

Flux, hosted by Deepgram, measures mean 99 ms time to final segment (9th of 29) and 5.8% word error rate (22nd of 31) among STT systems. Results cover the last 30 days. Coval benchmarks hosted endpoints daily using a consistent evaluation methodology. Last measured .

Flux General English is Deepgram's English streaming recognizer for conversational turn-taking.

Time to Final Segment#9 / 29
99ms
Word Error Rate#22 / 31
5.8%
Time to First Token#5 / 27
1091ms

Overview

Deepgram designed Flux around real-time voice-agent conversations and end-of-turn handling; this entry is the English-specific endpoint.

Flux 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
Deepgram
Hosted by
Deepgram
Source
Official API
Licensing
Proprietary
Deployment
On-prem
Region
US
Features
Keyterm biasing, VAD

How Flux 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 Flux 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 Pro168 ms
  3. #15Whisper Large v3via Together AI200 ms
  4. #16Pulse205 ms
  5. #17Grok STT223 ms
  6. #18resonant-1252 ms
  7. #19Default258 ms
  8. #20Linden 1271 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 1636 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 ms946 ms13,669
2STT RT v5Soniox45 ms1518 ms23,533
3Qwen3 ASR FastNari Labs47 ms1752 ms26,652
4STT 1Inworld AI66 ms1212 ms24,749
5Nova 3Deepgram85 ms1431 ms26,002
6Nova 2Deepgram87 ms1417 ms25,960
7Parakeet TDT 0.6B v3Together AI97 ms1228 ms26,460
8Flux MultilingualDeepgram98 ms1165 ms26,007
9FluxDeepgram99 ms1091 ms26,005
10Scribe v2 RealtimeElevenLabs122 ms2158 ms26,640
11Whisper Large v3Baseten131 ms924 ms13,663
12Universal 3.6 ProAssemblyAI135 ms1071 ms24,185
13Ink 2Cartesia148 ms1966 ms26,652
14Universal 3.5 ProAssemblyAI168 ms1025 ms26,654
15Whisper Large v3Together AI200 ms1261 ms26,445
16PulseSmallest205 ms1976 ms26,654
17Grok STTSpaceXAI223 ms—26,576
18resonant-1Reson8252 ms—26,653
19DefaultGradium258 ms1986 ms26,632
20Linden 1Speechmatics271 ms1537 ms26,595
21Gemini 3.5 Transcribe LiveGemini317 ms1609 ms26,567
22Daytona STTGuava399 ms5695 ms5,225
23Voxtral Mini Transcribe Realtime 2602Mistral410 ms1851 ms26,200
24GPT Realtime WhisperOpenAI599 ms1824 ms26,622
25Solaria 1Gladia636 ms1781 ms26,198
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 ms26,624
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: 5th of 27 on Time to First Token.

Latency vs accuracy

TTFS vs WEREach point is one measured model · Flux highlighted · 30-day averagesTime to Final Segment against Word Error Rate for every measured STT model, with Flux 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%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 — 148 ms, 3.6%Universal 3.5 Pro — 168 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 — 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 — 410 ms, 5.4%GPT Realtime Whisper — 599 ms, 4.0%Solaria 1 — 636 ms, 6.0%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%Flux — 99 ms, 5.8%

Where the errors come from

WER compositionFlux's Word Error Rate split by error type · 30-day averageFlux's WER split into substitutions, deletions and insertions.
  • Flux5.8%
SubstitutionsDeletionsInsertions

Averages and tail latency

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

Flux latency distributionMilliseconds · shared axis across metrics · last 30 daysFlux's latency percentiles per metric: p25–p75 band, p50 tick, whisker to max.
  • Time to Final Segmentp50 98 ms · Max 17313 ms
  • Time to First Tokenp50 1171 ms · Max 4623 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 Segment99 ms62 ms98 ms123 ms149 ms163 ms17313 ms26,005
Word Error Rate5.8%0.0%3.1%9.5%18.2%25.0%92.3%26,005
Time to First Token1091 ms775 ms1171 ms1254 ms1534 ms1975 ms4623 ms26,005

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 daysFlux's daily median Time to Final Segment over the last 30 days.
Flux · Sep 14: 101 msFlux · Sep 15: 100 msFlux · Sep 16: 112 msFlux · Sep 17: 86 msFlux · Sep 18: 107 msFlux · Sep 19: 99 msFlux · Sep 20: 90 msFlux · Sep 21: 92 msFlux · Sep 22: 105 msFlux · Sep 23: 92 msFlux · Sep 24: 98 msFlux · Sep 25: 114 msFlux · Sep 26: 88 msFlux · Sep 27: 98 msFlux · Sep 28: 102 msFlux · Sep 29: 107 msFlux · Sep 30: 72 msFlux · Oct 1: 107 msFlux · Oct 2: 90 msFlux · Oct 3: 98 msFlux · Oct 4: 98 msFlux · Oct 5: 113 msFlux · Oct 6: 99 msFlux · Oct 7: 82 msFlux · Oct 8: 101 msFlux · Oct 9: 114 msFlux · Oct 10: 77 msFlux · Oct 11: 92 msFlux · Oct 12: 88 msFlux · Oct 13: 115 msFlux · Oct 14: 108 ms
Word Error Rate — daily averageDaily average · UTC daysFlux's daily Word Error Rate over the last 30 days.
Flux · Sep 14: 6.9%Flux · Sep 15: 6.7%Flux · Sep 16: 7.0%Flux · Sep 17: 9.0%Flux · Sep 18: 5.9%Flux · Sep 19: 7.3%Flux · Sep 20: 6.3%Flux · Sep 21: 7.5%Flux · Sep 22: 6.7%Flux · Sep 23: 7.7%Flux · Sep 24: 6.8%Flux · Sep 25: 8.0%Flux · Sep 26: 7.9%Flux · Sep 27: 6.5%Flux · Sep 28: 6.1%Flux · Sep 29: 6.2%Flux · Sep 30: 7.0%Flux · Oct 1: 7.5%Flux · Oct 2: 5.8%Flux · Oct 3: 7.3%Flux · Oct 4: 7.3%Flux · Oct 5: 8.1%Flux · Oct 6: 7.5%Flux · Oct 7: 6.3%Flux · Oct 8: 7.5%Flux · Oct 9: 6.7%Flux · Oct 10: 6.4%Flux · Oct 11: 6.8%Flux · Oct 12: 6.6%Flux · Oct 13: 6.8%Flux · Oct 14: 11.4%

Time to Final Segment by dataset

Flux by test conditionMilliseconds · lower is better · best condition firstFlux's Time to Final Segment on each benchmark dataset.
  1. PipeCat (production)96 ms
  2. WildASR noise gaps100 ms
  3. WildASR far-field101 ms
  4. WildASR clean101 ms
  5. WildASR accents102 ms
  6. WildASR phone codec109 ms
  7. WildASR clipping112 ms
Time to Final Segment per dataset, with the number of samples behind each figure.
DatasetTTFSSamples
ProductionPipeCat96 ms13,678
AccentsWildASR102 ms1,369
CleanWildASR101 ms5,475
ClippingWildASR112 ms1,371
Far-fieldWildASR101 ms1,372
Noise gapsWildASR100 ms1,369
Phone codecWildASR109 ms1,371

Strongest condition: PipeCat (production) at 96 ms · weakest: WildASR clipping at 112 ms.

How fast is Flux?

On Deepgram, Flux measures mean 99 ms time to final segment (9th of 29) and mean 1091 ms time to first token (5th of 27). Last measured 2026-10-14.

How accurate is Flux?

On Deepgram, Flux measures 5.8% word error rate (22nd of 31). Last measured 2026-10-14.

Who hosts Flux?

Flux is created by Deepgram and served by Deepgram. Coval measures each hosted endpoint separately.

Limits of this comparison

Coval benchmarks it independently from Flux General Multi on fixed audio.

  • Coval's public metrics do not score every turn-detection event or application behavior supported by Flux.

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