Intelligent transcription with Gemini 3.5 Transcribe

Gemini 3.5 Transcribe is a speech-to-text model that converts raw audio into polished, formatted text while handling background noise, complex jargon, and self-corrections. It removes filler words like ums and ahs, auto-formats text, and supports custom vocabulary, multi-speaker identification for up to three speakers, and live language detection across more than 85 languages.

The model is available through two developer APIs. The Real-time streaming API uses the Live API for bidirectional streaming with sub-second latency, while the Pre-recorded audio processing API uses the Interactions API to transcribe recorded meetings and call logs with word-level timestamps and speaker attribution. According to Artificial Analysis, the model achieves an average Word Error Rate of 4.0 percent for streaming and 2.6 percent for non-streaming use cases. It also supports function calling to delegate tasks to other Gemini models.

Gemini 3.5 Transcribe is integrated into several consumer products and surfaces. It powers the Rambler feature on Gboard for Android, the Gemini app on macOS, Google Antigravity, and upcoming features in Chrome. Developers can access the model in public preview through Google AI Studio and the Gemini Enterprise Agent Platform.