Skip to content

SF·SIX / SOURCE-BACKED RESEARCH

Transcription models

Explore endpoints explicitly identified for transcription or automatic speech recognition.

7 models with current provider prices, most widely served first

Checked 2026-10-17 09:48 UTC · Source: fal.ai · Methodology

Models

Kotoba Speech-to-Text

kotoba-technologies · text

1 current source-qualified record

Cohere Transcribe

Unattributed (via fal) · text

1 current source-qualified record

Speech-to-Text (turbo)

Unattributed (via fal) · text

1 current source-qualified record

Speech-to-Text (turbo / stream)

Unattributed (via fal) · text

1 current source-qualified record

Speech-To-text (stream)

Unattributed (via fal) · text

1 current source-qualified record

Speech-to-Text

Unattributed (via fal) · text

1 current source-qualified record

Wizper (Whisper v3 -- fal.ai edition)

Unattributed (via fal) · text

1 current source-qualified record

Capabilities are source-reported and do not constitute an independent quality assessment. Community catalog references for this capability are listed in the interactive directory.