Speech-to-Text

Time from the last spoken word to the final transcript, decided by the model itself.

MetaMuse Voice TranscribeTry →301ms · 584ms

Sonioxstt-rt-v5Try →489ms · 2.40s

Sonioxstt-rt-v6555ms · 2.48s

AssemblyAIUniversal-3.6 ProTry →589ms · 1.58s

AssemblyAIUniversal-3.5 ProTry →606ms · 1.62s

DeepgramFlux MultilingualTry →654ms · 1.36s

NariQwen3-ASRTry →681ms · 960ms

NariQwen3-ASR FastTry →684ms · 965ms

Omimedical-1768ms · 1.04s

DeepgramNova-3Try →771ms · 1.21s

DeepgramFluxTry →773ms · 1.48s

ElevenLabsScribe v2 RealtimeTry →793ms · 1.06s

Hoppernemotron-asrTry →809ms · 1.08s

OpenAIGPT-4o mini TranscribeTry →926ms · 1.35s

Smallest AIPulseTry →958ms · 1.45s

GladiaSolaria-1Try →1.09s · 1.59s

Fish Audiotranscribe-stream-11.12s · 2.27s

GradiumASRTry →1.22s · 1.35s

ModulateVelma 2Try →1.25s · 1.54s

OpenAIGPT-4o TranscribeTry →1.28s · 1.70s

MistralVoxtral Mini Transcribe Realtime 31.28s · 3.34s

xAIGrok Voice Transcribe 2.01.35s · 1.75s

xAIGrok Voice Transcribe 1.0Try →1.40s · 1.95s

SpeechmaticsStandard1.45s · 1.81s

InworldRealtime STT-11.90s · 2.71s

0ms500ms1.00s1.50s2.00s2.50s

p50 p90

Time from the caller starting to speak to the first word of the live transcript.

AzureMAI-Transcribe-2 Streaming241ms · 342ms

Omimedical-1309ms · 550ms

Fish Audiotranscribe-stream-1414ms · 480ms

DeepgramFlux MultilingualTry →506ms · 1.65s

DeepgramFluxTry →641ms · 1.19s

GladiaSolaria-1Try →671ms · 1.20s

SpeechmaticsStandard690ms · 1.05s

MetaMuse Voice TranscribeTry →728ms · 1.16s

AssemblyAIUniversal-3.6 ProTry →775ms · 974ms

AssemblyAIUniversal-3.5 ProTry →775ms · 973ms

Sonioxstt-rt-v5Try →786ms · 1.27s

Hoppernemotron-asrTry →825ms · 1.67s

DeepgramNova-3Try →885ms · 1.45s

InworldRealtime STT-1942ms · 1.02s

MistralVoxtral Mini Transcribe Realtime 3978ms · 1.61s

OpenAIGPT Realtime Whisper987ms · 1.27s

OpenAIGPT Live TranscribeTry →988ms · 1.13s

xAIGrok Voice Transcribe 1.0Try →989ms · 1.10s

xAIGrok Voice Transcribe 2.0995ms · 1.18s

Smallest AIPulseTry →1.04s · 2.01s

Sonioxstt-rt-v61.11s · 1.60s

CartesiaInk-2Try →1.19s · 1.61s

ModulateVelma 2Try →1.38s · 1.47s

NariQwen3-ASR FastTry →1.65s · 1.74s

GradiumASRTry →1.65s · 2.43s

NariQwen3-ASRTry →1.68s · 1.77s

ElevenLabsScribe v2 RealtimeTry →2.02s · 2.19s

OpenAIGPT-4o mini TranscribeTry →2.91s · 6.12s

OpenAIGPT-4o TranscribeTry →4.72s · 8.46s

0ms1.00s2.00s3.00s4.00s5.00s

p50 p90

Time from a word being spoken to it appearing in the live transcript for good.

AzureMAI-Transcribe-2 Streaming114ms · 515ms

DeepgramFlux MultilingualTry →180ms · 715ms

DeepgramFluxTry →190ms · 586ms

Fish Audiotranscribe-stream-1249ms · 1.57s

Omimedical-1343ms · 2.76s

SpeechmaticsStandard403ms · 1.07s

Hoppernemotron-asrTry →425ms · 717ms

MetaMuse Voice TranscribeTry →486ms · 716ms

Sonioxstt-rt-v5Try →496ms · 739ms

InworldRealtime STT-1583ms · 1.86s

DeepgramNova-3Try →583ms · 1.72s

AssemblyAIUniversal-3.5 ProTry →803ms · 2.52s

MistralVoxtral Mini Transcribe Realtime 3804ms · 947ms

AssemblyAIUniversal-3.6 ProTry →815ms · 2.71s

Sonioxstt-rt-v6857ms · 1.11s

ModulateVelma 2Try →896ms · 2.35s

NariQwen3-ASR FastTry →901ms · 2.04s

ElevenLabsScribe v2 RealtimeTry →945ms · 1.97s

NariQwen3-ASRTry →973ms · 2.18s

Smallest AIPulseTry →1.17s · 1.98s

CartesiaInk-2Try →1.22s · 1.71s

GradiumASRTry →1.48s · 1.66s

GladiaSolaria-1Try →1.50s · 3.98s

xAIGrok Voice Transcribe 2.01.89s · 10.22s

xAIGrok Voice Transcribe 1.0Try →2.08s · 8.91s

OpenAIGPT-4o mini TranscribeTry →2.47s · 5.29s

OpenAIGPT-4o TranscribeTry →3.43s · 7.18s

OpenAIGPT Realtime Whisper3.96s · 11.28s

OpenAIGPT Live TranscribeTry →4.02s · 11.34s

0ms1.00s2.00s3.00s4.00s5.00s

p50 p90

Words the live transcript takes back, per word in the final transcript.

GradiumASRTry →0%

CartesiaInk-2Try →0%

MetaMuse Voice TranscribeTry →0%

MistralVoxtral Mini Transcribe Realtime 30%

Smallest AIPulseTry →0%

OpenAIGPT-4o TranscribeTry →3%

OpenAIGPT-4o mini TranscribeTry →5%

OpenAIGPT Realtime Whisper6%

Hoppernemotron-asrTry →8%

OpenAIGPT Live TranscribeTry →10%

Sonioxstt-rt-v610%

Sonioxstt-rt-v5Try →13%

NariQwen3-ASRTry →21%

NariQwen3-ASR FastTry →22%

ElevenLabsScribe v2 RealtimeTry →29%

InworldRealtime STT-130%

DeepgramNova-3Try →33%

ModulateVelma 2Try →33%

SpeechmaticsStandard35%

AssemblyAIUniversal-3.6 ProTry →35%

AssemblyAIUniversal-3.5 ProTry →36%

AzureMAI-Transcribe-2 Streaming37%

DeepgramFluxTry →40%

DeepgramFlux MultilingualTry →41%

Fish Audiotranscribe-stream-156%

xAIGrok Voice Transcribe 2.064%

xAIGrok Voice Transcribe 1.0Try →83%

Omimedical-189%

GladiaSolaria-1Try →156%

0%40%80%120%160%

Published list price per thousand minutes; batch-only models show their batch rate.

$0.00$5.00$10.00$15.00$20.00

$0.83

$0.83

$1.00

$1.67

$1.67

$2.00

$2.00

$2.00

$2.00

$2.10

$2.50

$2.50

$2.50

$2.67

$3.00

$3.00

$3.30

$3.30

$3.50

$3.70

$4.00

$4.00

$4.00

$4.50

$4.80

$6.00

$6.00

$6.00

$6.50

$6.50

$7.17

$7.50

$7.50

$7.50

$7.80

$8.00

$9.00

$10.40

$12.50

$17.00

$17.00

ModulateVelma 2Hoppernemotron-asrNariQwen3-ASRSonioxstt-async-v6AzureMAI-Transcribe-2NariQwen3-ASR FastSonioxstt-rt-v6Sonioxstt-rt-v5HopperNemotron 3.5 TranscribeAlibabaQwen3-ASR FlashInworldRealtime STT-1AssemblyAIUniversal-Streaming EnglishAssemblyAIUniversal-Streaming MultilingualSpeechmaticsLinden 1MetaMuse Voice TranscribeOpenAIGPT-4o mini TranscribexAIGrok Voice Transcribe 1.0xAIGrok Voice Transcribe 2.0MistralVoxtral Mini Transcribe 3ElevenLabsScribe v2Smallest AIPulseSpeechmaticsStandardMistralVoxtral Small Transcribe 3OpenAIGPT TranscribeDeepgramNova-3Fish Audiotranscribe-stream-1OpenAIGPT-4o TranscribeFish AudioTranscribe 1 ProElevenLabsScribe v2 RealtimeDeepgramFluxSpeechmaticsEnhancedAssemblyAIUniversal-3.6 ProAssemblyAIUniversal-3.5 ProOmimedical-1DeepgramFlux MultilingualMistralVoxtral Mini Transcribe Realtime 3CartesiaInk-2GradiumASRGladiaSolaria-1OpenAIGPT Live TranscribeOpenAIGPT Realtime Whisper