Google has announced Gemini 3.5 Transcribe, an AI-powered speech-to-text model that cleans up verbal filler words like "ums" and corrects itself in real time. It claims to be 70% faster than its predecessor Chirp 3, with live-speech error rate dropping to 5.5%, and supports 85 languages with up to three speakers.
Background
Google has been steadily improving its speech recognition technology, previously introducing the Gboard
- Source
- Ars Technica
- Published
- Aug 27, 2026 at 03:19 AM
- Score
- 6.0 / 10