Google introduced Gemini 3.5 Live Translate on June 9, 2026, a new AI model designed to facilitate real-time, natural-sounding translation across more than 70 languages. The technology eliminates the friction of language barriers, allowing users to engage in fluid conversations while preserving original intonation, tempo, and pitch.
The primary shift with Gemini 3.5 Live Translate is its ability to generate continuous translated speech. Unlike previous iterations that required a speaker to finish a full sentence before providing a translation, this model processes and outputs audio a few seconds behind the speaker. This creates a more natural flow, making it suitable for high-stakes environments like international meetings or spontaneous travel interactions.
The update provides significant improvements over previous Google Meet speech translation limits, which were restricted to only 5 languages. The new capabilities include support for over 70 languages, more than 2,000 language combinations in a single meeting, and natural speech synthesis that retains original speaker nuances.
Google is rolling out the update across several platforms, though availability varies depending on device and professional needs. The Google Translate app is currently rolling out worldwide for both Android and iOS devices. Android Listening Mode streams translated audio directly to the earpiece for hands-free use. Google Meet will begin a private preview for select Google Workspace enterprise customers in June 2026, with a broader release planned later this year.
To address growing concerns regarding AI-generated content, Google has integrated SynthID watermarking into all audio produced by Gemini 3.5 Live Translate. This ensures that any audio generated by the model can be identified as AI-produced, maintaining transparency in both professional and personal settings.
Developers can already access the model in preview mode via the Gemini Live API and Google AI Studio. These platforms provide demo code and examples to help integrate real-time translation into custom workflows.
