Alphabet's Google introduced Gemini 3.5 Transcribe, its most advanced speech-to-text AI model. It converts raw audio into polished, formatted text. The model handles background noise, complex jargon, and verbal disfluencies. It detects and transcribes over 85 languages. The model identifies up to three speakers in a recording.
Compared to its predecessor, Chirp 3, the new model reduces final transcription time by 70%. Google rolls out Gemini 3.5 Transcribe for developers and enterprises through its AI platforms. It integrates into consumer products like the Gemini app on macOS. The model also integrates into the Rambler feature on Android. A future update will bring its capabilities to the Chrome browser. This will allow users to dictate text into any web field.