Voice AI BenchmarksBy

PROVIDERLAST 30 DAYS

official resource

NVIDIA voice AI models and benchmarks

NVIDIA lists 2 STT models in Coval. Its fastest dated STT result is Parakeet TDT 0.6B v3 on Together AI at mean 97 ms time to final segment (7th of 29) among STT systems, with 11.1% WER. Results cover the last 30 days. Last measured .

NVIDIA creates the open-weight Nemotron Speech Streaming and Parakeet TDT speech recognition models.

Measured models
2
STT
Best STT model#7 / 29
97ms
Parakeet TDT 0.6B v3

Overview

Nemotron Speech Streaming and Parakeet TDT expose model-card and weight identities while Together supplies the live serving path.

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
Nemotron 3.5 ASR Streaming, Parakeet TDT 0.6B v3

Speech-to-Text

Full STT dashboard

Ranked on Time to Final Segment against 29 measured models.

Time to Final Segmentms · lower is better · NVIDIA models markedEvery measured STT model on Time to Final Segment, with NVIDIA's models highlighted.
  1. #1Qwen3 ASR 1.7b39 ms
  2. #2STT RT v544 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
leaders plus NVIDIA models · 22 other models in the full table
Word Error Rate% · lower is better · NVIDIA models markedEvery measured STT model on Word Error Rate, with NVIDIA's models highlighted.
  1. #1Universal 3.6 Pro2.1%
  2. #2Universal 3.5 Pro2.4%
  3. #3Qwen3 ASR Fast2.6%
  4. #4resonant-12.6%
  5. #5Qwen3 ASR 1.7b2.9%
  6. #6Gemini 3.5 Transcribe Live3.0%
  7. #7Chirp 33.2%
  8. #30Parakeet TDT 0.6B v311.1%
  9. #31Nemotron 3.5 ASR Streaming13.2%
leaders plus NVIDIA models · 22 other models in the full table
Time to First Tokenms · lower is better · NVIDIA models markedEvery measured STT model on Time to First Token, with NVIDIA's models highlighted.
  1. #1Whisper Large v3via Baseten923 ms
  2. #2Qwen3 ASR 1.7b945 ms
  3. #3Universal 3.5 Pro1024 ms
  4. #4Universal 3.6 Pro1071 ms
  5. #5Flux1091 ms
  6. #6Flux Multilingual1165 ms
  7. #7STT 11188 ms
  8. #8Parakeet TDT 0.6B v31224 ms
  9. #15Nemotron 3.5 ASR Streaming1544 ms
leaders plus NVIDIA models · 18 other models in the full table
TTFS vs WER — NVIDIA vs every measured modelEach point is one measured STT model · NVIDIA models highlighted · 30-day averagesTime to Final Segment against Word Error Rate for every measured STT model, with NVIDIA's models highlighted.
Qwen3 ASR 1.7b — 39 ms, 2.9%STT RT v5 — 44 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%Flux Multilingual — 98 ms, 7.5%Flux — 98 ms, 5.8%Scribe v2 Realtime — 121 ms, 4.3%Whisper Large v3 via Baseten — 131 ms, 4.4%Universal 3.6 Pro — 134 ms, 2.1%Ink 2 — 149 ms, 3.6%Universal 3.5 Pro — 168 ms, 2.4%Whisper Large v3 via Together AI — 196 ms, 7.4%Pulse — 205 ms, 4.6%Grok STT — 222 ms, 4.1%resonant-1 — 251 ms, 2.6%Default — 262 ms, 8.7%Linden 1 — 275 ms, 3.9%Gemini 3.5 Transcribe Live — 320 ms, 3.0%Daytona STT — 399 ms, 4.9%Voxtral Mini Transcribe Realtime 2602 — 419 ms, 5.3%Solaria 1 — 577 ms, 6.1%GPT Realtime Whisper — 599 ms, 4.1%GPT-4o mini Transcribe — 711 ms, 3.9%Chirp 3 — 766 ms, 3.2%GPT-4o Transcribe — 784 ms, 3.7%Chirp 2 — 910 ms, 3.9%Parakeet TDT 0.6B v3 — 97 ms, 11.1%
Benchmarked models with their Time to Final Segment over the last 30 days.
ModelHostTTFSRank
Nemotron 3.5 ASR StreamingTogether AI——
Parakeet TDT 0.6B v3Together AI97 ms7th of 29

How fast are NVIDIA's STT models?

Parakeet TDT 0.6B v3 on Together AI measures mean 97 ms time to final segment (7th of 29) among STT systems. Last measured 2026-10-15.

How accurate are NVIDIA's STT models?

Parakeet TDT 0.6B v3 on Together AI measures 11.1% word error rate (30th of 31) among STT systems. Last measured 2026-10-15. Nemotron 3.5 ASR Streaming on Together AI measures 13.2% word error rate (31st of 31) among STT systems. Last measured 2026-10-15.

Which NVIDIA model is fastest?

Its fastest dated STT result is Parakeet TDT 0.6B v3 on Together AI at mean 97 ms time to final segment (7th of 29) among STT systems, with 11.1% WER. Last measured 2026-10-15.

Limits of this comparison

Together hosts the endpoints measured by Coval.

  • Latency from Together should not be generalized to NVIDIA NIM, self-hosting or other optimized deployments of the same weights.

Official resources

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