Voice AI BenchmarksBy

PROVIDERLAST 30 DAYS

Nari Labs voice AI models and benchmarks

Nari Labs lists 2 STT and TTS models in Coval. Fastest dated mean latency over 30 days: STT: Qwen3 ASR Fast at 50 ms TTFS, with 2.6% WER. TTS: Qwen3 TTS Fast at 68 ms TTFA, with 1.8% WER. Last measured .

Coval measures voice models created or hosted by Nari Labs.

Measured models
2
STTTTS
Best STT model#3 / 28
50ms
Qwen3 ASR Fast
Best TTS model#3 / 32
68ms
Qwen3 TTS Fast

Overview

Every model below is measured daily on the same fixed inputs and ranked against the full field, never blended into a company score.

Model lineup

Speech-to-Text
serves Qwen3 ASR Fast
Text-to-Speech
serves Qwen3 TTS Fast

Speech-to-Text

Full STT dashboard

Ranked on Time to Final Segment against 28 measured models.

Time to Final Segmentms · lower is better · Nari Labs models markedEvery measured STT model on Time to Final Segment, with Nari Labs's models highlighted.
  1. #1Qwen3 ASR 1.7b35 ms
  2. #2STT RT v544 ms
  3. #3Qwen3 ASR Fast50 ms
  4. #4STT 170 ms
  5. #5Nova 385 ms
  6. #6Nova 287 ms
  7. #7Parakeet TDT 0.6B v388 ms
leaders plus Nari Labs models · 21 other models in the full table
Word Error Rate% · lower is better · Nari Labs models markedEvery measured STT model on Word Error Rate, with Nari Labs's models highlighted.
  1. #1Universal 3.6 Pro2.2%
  2. #2Universal 3.5 Pro2.4%
  3. #3Qwen3 ASR Fast2.6%
  4. #4Realtime Global2.6%
  5. #5resonant-12.6%
  6. #6Gemini 3.5 Transcribe Live2.9%
  7. #7Qwen3 ASR 1.7b2.9%
leaders plus Nari Labs models · 23 other models in the full table
Time to First Tokenms · lower is better · Nari Labs models markedEvery measured STT model on Time to First Token, with Nari Labs's models highlighted.
  1. #1Whisper Large v3via Baseten914 ms
  2. #2Qwen3 ASR 1.7b936 ms
  3. #3Universal 3.5 Pro1019 ms
  4. #4STT 11026 ms
  5. #5Universal 3.6 Pro1080 ms
  6. #6Flux1097 ms
  7. #7Flux Multilingual1174 ms
  8. #18Qwen3 ASR Fast1755 ms
leaders plus Nari Labs models · 20 other models in the full table
TTFS vs WER — Nari Labs vs every measured modelEach point is one measured STT model · Nari Labs models highlighted · 30-day averagesTime to Final Segment against Word Error Rate for every measured STT model, with Nari Labs's models highlighted.
Qwen3 ASR 1.7b — 35 ms, 2.9%STT RT v5 — 44 ms, 4.8%STT 1 — 70 ms, 4.4%Nova 3 — 85 ms, 5.2%Nova 2 — 87 ms, 6.6%Parakeet TDT 0.6B v3 — 88 ms, 10.9%Flux Multilingual — 100 ms, 7.4%Flux — 101 ms, 5.7%Universal 3.6 Pro — 114 ms, 2.2%Scribe v2 Realtime — 118 ms, 4.2%Universal 3.5 Pro — 126 ms, 2.4%Whisper Large v3 via Baseten — 126 ms, 4.5%Ink 2 — 146 ms, 3.4%Realtime Global — 159 ms, 2.6%Whisper Large v3 via Together AI — 175 ms, 7.2%Grok STT — 180 ms, 4.1%Pulse — 203 ms, 4.6%resonant-1 — 243 ms, 2.6%Linden 1 — 254 ms, 3.8%Default — 261 ms, 8.4%Gemini 3.5 Transcribe Live — 383 ms, 2.9%Daytona STT — 384 ms, 4.8%Solaria 1 — 414 ms, 6.2%Voxtral Mini Transcribe Realtime 2602 — 436 ms, 5.3%GPT Realtime Whisper — 563 ms, 4.0%Chirp 3 — 807 ms, 3.2%Chirp 2 — 896 ms, 3.9%Qwen3 ASR Fast — 50 ms, 2.6%
Benchmarked models with their Time to Final Segment over the last 30 days.
ModelHostTTFSRank
Qwen3 ASR FastNari Labs50 ms3rd of 28

Text-to-Speech

Full TTS dashboard

Ranked on Time to First Audio against 32 measured models.

Time to First Audioms · lower is better · Nari Labs models markedEvery measured TTS model on Time to First Audio, with Nari Labs's models highlighted.
  1. #1vui42 ms
  2. #2Default64 ms
  3. #3Qwen3 TTS Fast68 ms
  4. #4TTS Flash 283 ms
  5. #5Qwen3 TTS 1.7b105 ms
  6. #6Simba 3.0110 ms
  7. #7Simba 3.2118 ms
leaders plus Nari Labs models · 25 other models in the full table
Word Error Rate% · lower is better · Nari Labs models markedEvery measured TTS model on Word Error Rate, with Nari Labs's models highlighted.
  1. #1Sonic 3.61.5%
  2. #2Sonic 3.51.5%
  3. #3Lightning v3.1 Pro1.6%
  4. #4S2.1 Pro Free1.6%
  5. #5S2.1 Pro1.6%
  6. #6Simba 3.01.6%
  7. #7Assemblyai TTS 11.6%
  8. #18Qwen3 TTS Fast1.8%
leaders plus Nari Labs models · 24 other models in the full table
TTFA vs WER — Nari Labs vs every measured modelEach point is one measured TTS model · Nari Labs models highlighted · 30-day averagesTime to First Audio against Word Error Rate for every measured TTS model, with Nari Labs's models highlighted.
vui — 42 ms, 2.0%Default — 64 ms, 1.9%TTS Flash 2 — 83 ms, 1.7%Qwen3 TTS 1.7b — 105 ms, 1.9%Simba 3.0 — 110 ms, 1.6%Simba 3.2 — 118 ms, 2.0%Palabra TTS — 123 ms, 2.8%PhantomZ 4.0 — 141 ms, 2.2%TTS 2 — 163 ms, 1.7%drama3-preview — 165 ms, 1.6%Falcon 2 — 177 ms, 1.6%Eleven v4 Turbo — 194 ms, 1.7%Flash v2.5 — 202 ms, 1.9%Flux TTS — 207 ms, 1.8%Assemblyai TTS 1 — 223 ms, 1.6%TTS RT v1 — 261 ms, 2.6%TTS Rt v2 — 264 ms, 2.7%Airy TTS — 281 ms, 1.6%Sonic 3.6 — 287 ms, 1.5%Sonic 3.5 — 291 ms, 1.5%Aura 2 — 303 ms, 1.9%Grok TTS — 342 ms, 1.7%S2.1 Pro — 367 ms, 1.6%Coda — 377 ms, 2.4%Lightning v3.1 Pro — 383 ms, 1.6%Mist v3 — 387 ms, 2.0%Daytona TTS — 439 ms, 1.8%Aura 2 En — 546 ms, 2.1%Chirp 3 HD — 588 ms, 1.6%Qwen3 TTS Flash Realtime — 727 ms, 3.0%S2.1 Pro Free — 802 ms, 1.6%Qwen3 TTS Fast — 68 ms, 1.8%
Benchmarked models with their Time to First Audio over the last 30 days.
ModelHostTTFARank
Qwen3 TTS FastNari Labs68 ms3rd of 32

How fast are Nari Labs's STT and TTS models?

Qwen3 ASR Fast measures mean 50 ms time to final segment (3rd of 28) among STT systems. Last measured 2026-11-02. Qwen3 TTS Fast measures mean 68 ms time to first audio (3rd of 32) among TTS systems. Last measured 2026-11-02.

How accurate are Nari Labs's STT and TTS models?

Qwen3 ASR Fast measures 2.6% word error rate (3rd of 30) among STT systems. Last measured 2026-11-02. Qwen3 TTS Fast measures 1.8% word error rate (18th of 32) among TTS systems. Last measured 2026-11-02.

Which Nari Labs model is fastest?

Its fastest dated STT result is Qwen3 ASR Fast at mean 50 ms time to final segment (3rd of 28) among STT systems, with 2.6% WER. Last measured 2026-11-02. Its fastest dated TTS result is Qwen3 TTS Fast at mean 68 ms time to first audio (3rd of 32) among TTS systems, with 1.8% WER. Last measured 2026-11-02.

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