Catalog · 40 models · 13 providers

Transcription models.

Every speech-to-text model worth using, one rate card.

Total models

106

Languages

$0.0107

Cheapest /min

9.4%

Best WER

6 models

filtered from 40

Chirp 3

Google Cloud

Batch

United States

Google's latest generative ASR foundation model — 85+ languages

$0.0107/min

WER

9.37%

CER

4.79%

Speed Factor

0.3x

Uptime

100.0%

Billing

per 15s

Retention

Custom

72 languages

(view all)

GenerativeAuto-detectDiarization+2

Transcribe View Benchmarks

Google Command & Search

Google Cloud

Batch

United States

Optimized for short queries and voice commands

$0.016/min

WER

47.19%

CER

21.87%

Speed Factor

0.1x

Billing

per 15s

Retention

Custom

40 languages

(view all)

Word Timestamps

Transcribe View Benchmarks

Google Default

Google Cloud

Batch

United States

General-purpose model

$0.016/min

WER

45.74%

CER

36.56%

Speed Factor

0.2x

Billing

per 15s

Retention

Custom

40 languages

(view all)

DiarizationWord Timestamps

Transcribe View Benchmarks

Google Latest (Long)

Google Cloud

Batch

United States

Conformer model for long-form audio (minutes to hours)

$0.0107/min

WER

22.54%

CER

10.26%

Speed Factor

0.4x

Billing

per 15s

Retention

Custom

40 languages

(view all)

Word Timestamps

Transcribe View Benchmarks

Google Latest (Short)

Google Cloud

Batch

United States

Conformer model for short utterances (< 60s)

$0.016/min

WER

63.82%

CER

57.52%

Speed Factor

0.1x

Billing

per 15s

Retention

Custom

40 languages

(view all)

Word Timestamps

Transcribe View Benchmarks

Google Telephony

Google Cloud

Batch

United States

Optimized for telephony audio (8kHz)

$0.016/min

WER

22.85%

CER

13.22%

Speed Factor

0.2x

Uptime

100.0%

Billing

per 15s

Retention

Custom

9 languages

(view all)

Word Timestamps

Transcribe View Benchmarks