Google Launches Gemini 3.5 Transcribe

Google has released Gemini 3.5 Transcribe, an AI model designed to convert unstructured speech into formatted text. The model automatically removes filler words, handles self-corrections, and supports over 85 languages. It is currently available for developers through Google AI Studio and the Gemini Enterprise Agent Platform.

The model achieves a 2.6% word error rate for non-streaming applications and a 4.0% rate for streaming. It supports speaker attribution for up to three individuals.