Gemini 3.5 Transcribe is Google's new AI transcription model that removes disfluencies like 'ums' and 'ahs'. It supports 85+ languages, detects specialized jargon, and improves on the previous Chirp 3 model for multilingual accuracy and word error rate. The model also lets users edit transcripts naturally by voice, expanding Google's Gemini Audio lineup.
Opening Kapyn…