Cohere Transcribe (03-2026)

Cohere Transcribe (March 2026) is a speech-to-text model from Cohere, converting audio recordings into accurate text transcriptions. It supports multiple languages and handles various audio qualities, accents, and speaking styles.

The model provides timestamped transcriptions with speaker diarization capabilities, making it suitable for meeting transcription, podcast processing, and customer call analysis. It integrates with Cohere's broader AI platform for end-to-end voice-to-insight pipelines.

Transcribe v3 is designed for enterprise applications requiring reliable, scalable speech recognition across diverse audio sources.