Fluid, natural voice translation with Gemini 3.5 Live Translate

Gemini 3.5 Live Translate is a new audio model designed for near real-time, speech-to-speech translation across more than 70 languages. Unlike traditional systems that wait for a speaker to finish before translating, this model processes speech continuously as it is streamed. It preserves the original speaker's intonation, pacing, and pitch, resulting in fluid audio that remains only a few seconds behind the live conversation.

The model is built to handle noisy, unpredictable environments and automatically detects languages without requiring manual configuration. To ensure transparency and safety, all audio generated by the model is embedded with an imperceptible SynthID watermark, which helps identify AI-generated content.

Google is making this technology available through several channels. Developers can access it via the Gemini Live API and Google AI Studio, while enterprise users will see it integrated into Google Meet to support multilingual meetings with over 2,000 language combinations. For general users, the feature is rolling out to the Google Translate app on Android and iOS, including a new listening mode that allows users to hear translations privately through their phone’s earpiece. Early adopters, such as the ride-hailing service Grab, are already using the technology to facilitate real-time communication between drivers and passengers.