Intelligent transcription with Gemini 3.5 Transcribe

Gemini 3.5 Transcribe is a new speech-to-text model designed to convert raw audio directly into polished, formatted text while handling background noise, complex jargon, and natural disfluencies. The model is available for developers through the Gemini API in Google AI Studio and the Gemini Enterprise Agent Platform. It operates across two separate application programming interfaces, providing a real-time streaming option with sub-second latency via the Live API and a pre-recorded audio processing option that includes speaker attribution and word-level timestamps via the Interactions API.

The model introduces several smart transcription capabilities, including the removal of filler words, the automatic formatting of text, and the handling of self-corrections. According to Artificial Analysis measurements, it achieves an average word error rate of 4.0 percent for streaming and 2.6 percent for non-streaming use cases, while automatically detecting and transcribing over 85 languages with regional accent and custom vocabulary support. It also features function calling, allowing it to delegate complex tasks to other Gemini models. The model improves significantly over the previous Chirp 3 model, reducing time to final transcription by 70 percent.

Beyond developer tools, Gemini 3.5 Transcribe integrates into everyday Google surfaces and applications. Consumers can use it through the new Rambler feature on Android Gboard, the Gemini app on macOS, Google Antigravity, and soon on Google Chrome. These integrations allow users to dictate text, make inline voice edits, utilize screen context for transcription accuracy, and execute voice commands that trigger image generation and file analysis.