Google's Gemini 3.5 Transcribe turns speech to text in 85 languages while auto-correcting your verbal stumbles

Google launched Gemini 3.5 Transcribe, a speech-to-text model for real-time and recorded audio.

It supports automatic language detection in more than 85 languages and cleans transcripts by removing filler words, correcting spoken mistakes, and formatting text.

The low-latency model expands Google’s multilingual speech AI for developers and products including Gboard, Gemini, Chrome, and enterprise tools.