Google opens Gemini 3.5 Live Translate API to developers

Google Launches Gemini 3.5 Live Translate, Bringing Real-Time Speech Translation to 70+ Languages

Google on Monday unveiled Gemini 3.5 Live Translate, a speech-to-speech translation model that delivers near real-time audio translation across more than 70 languages. The model generates translated speech continuously as a person talks, rather than waiting for them to finish speaking, and preserves the speaker’s intonation, pacing, and pitch.

The launch marks the culmination of a phased rollout that began with a beta on Android in December 2025 and expanded to iOS in March 2026. Gemini 3.5 Live Translate is now available to developers through the Gemini Live API and Google AI Studio in public preview.youtube

From Headphones to the Boardroom

The model is rolling out across multiple Google products. On the consumer side, it powers the Live Translate feature in the Google Translate app on both Android and iOS, turning any pair of headphones into a real-time translation device.macrumors

For enterprise users, Google is bringing speech translation powered by Gemini 3.5 Live Translate to Google Meet in private preview for select Workspace business customers. The update expands Meet’s previous speech translation capability, which was limited to a handful of languages, to support over 70 languages and more than 2,000 language pairs in a single meeting. The model automatically detects which language a participant is speaking and translates it into each listener’s preferred language without requiring manual settings adjustments.investing

How It Works

Thor Schaeff and Anuda Weerasinghe from Google DeepMind introduced the model in a video published Monday, explaining that Gemini 3.5 Live Translate processes speech in a streaming fashion rather than in discrete chunks. The system automatically detects languages and filters out background noise while maintaining a natural listening experience.youtube

Google highlighted a partnership with Grab, the Southeast Asian ride-hailing company, which has been testing the model to improve communication between its users across different languages.youtube

A Year in the Making

Google first previewed its Gemini-powered live speech-to-speech translation capabilities in December 2025, when it launched the headphone-based beta on Android in the United States, Mexico, and India. At the time, the company said it would bring the feature to iOS and additional countries in 2026, along with broader API access. The iOS expansion arrived in March, and Monday’s announcement delivers on the remaining commitments with full developer API access and the Google Meet integration.cryptonomist

A broader rollout of the Meet feature to more Workspace customers is expected later this year.youtube