Google has announced 'Gemini 3.5 Live Translate,' which enables instant translation, and will be implemented in the Google Translate app for iOS and Android.

On June 9, 2026, Google announced its voice translation model, ' Gemini 3.5 Live Translate .' Supporting over 70 languages, including Japanese, this model translates speech into another language in near real-time while preserving the speaker's intonation, speaking speed, and pitch.

Gemini 3.5 Live Translate is here
https://blog.google/innovation-and-ai/models-and-research/gemini-models/gemini-live-3-5-translate/

Gemini 3.5 Live Translate is a model that continuously processes the audio stream, rather than the conventional method of waiting for the utterance to finish before translating. It features the generation of translated audio a few seconds after utterance to take context into account, and can automatically detect over 70 languages and generate natural-sounding speech that preserves the speaker's intonation, pauses, and pitch.

You can see Gemini 3.5 Live Translate translating English to Japanese in real time, starting around 1 minute and 39 seconds into the following video released by Google.

Introducing Gemini 3.5 Live Translate - YouTube


Gemini 3.5 Live Translate is available to developers as a public preview through the Gemini Live API and Google AI Studio. It handles multilingual input without manual configuration and features noise tolerance to handle noisy and unpredictable environments.

Google is promoting the Gemini Live API, which allows developers to build live interpretation features for multilingual calls, conferences, classes, and broadcasts. As of the time of writing, the Gemini Live API is already providing the complex infrastructure for real-time media delivery on developer platforms such as Agora, Fishjam, LiveKit, Pipecat, and Vision Agents.

Grab, the ride-hailing service provider, is piloting Gemini 3.5 Live Translate for multilingual communication between drivers and travelers during rides. Grab users reportedly make more than 10 million voice calls per month through the service, and Philip Kandal, Grab's Chief Product Officer, praises the automatic multilingual detection and accurate, low-latency voice translation.

Fluid, natural voice translation with Gemini 3.5 live Translate - YouTube


Furthermore, Google Meet, the audio conferencing tool, will be introducing voice translation using Gemini 3.5 Live Translate. The number of supported languages will expand from the current 5 to over 70, and while it was previously limited to translation between English and other languages, it will now support over 2000 language combinations in a single meeting. This Google Meet feature will be launched as a private preview for select Google Workspace corporate users starting in June 2026. A wider rollout is planned for the latter half of 2026.

Speech translation in Google Meet with Gemini 3.5 Live Translate - YouTube


The Google Translate app will implement Gemini 3.5 Live Translate on both Android and iOS. With the live translation feature, connecting headphones will allow users to access translations that reflect the speaker's tone in over 70 languages. Furthermore, a new 'Listening Mode' will be introduced for Android. This allows users to hear the translated audio directly through the earpiece by holding their smartphone to their ear, just like a normal phone call, even if headphones are not readily available.

Furthermore, all audio generated by Gemini 3.5 Live Translate will incorporate a SynthID watermark. Google states that this invisible watermark is directly embedded in the audio output, making AI-generated content detectable and helping to prevent misinformation.

Gemini 3.5 Live Translate will be rolled out starting June 9, 2026. It will be available in public preview for developers via the Gemini Live API and Google AI Studio, in private preview for select Google Workspace customers via Google Meet starting the same month, and globally for general users via the Android and iOS versions of the Google Translate app.