Gemini 3.5 Live Translate

Key Features

• •Continuous Speech Translation: Gemini 3.5 Live Translate generates translated speech continuously as the speaker talks, rather than waiting for pauses. This eliminates awkward gaps and preserves the speaker's natural intonation, pacing, and pitch, enabling fluid, near real-time conversations across languages.
• •Automatic Language Detection: The model automatically detects over 70 languages from the input audio, removing the need to manually select source languages. This simplifies usage in multilingual settings where participants may switch languages, ensuring seamless translation without interruptions for configuration.
• •Noise Robustness: Gemini 3.5 Live Translate is designed to handle loud and unpredictable environments. Its noise robustness ensures high translation accuracy and clear audio output even in crowded or noisy settings, making it suitable for real-world applications like meetings and public events.
• •Developer Platform Integrations: The model integrates with major developer platforms including Agora, Fishjam, LiveKit, Pipecat, and Vision Agents. These partnerships handle real-time media streaming infrastructure, allowing developers to quickly build and deploy voice translation apps using the Gemini Live API.
• •SynthID Watermarking: All audio generated by Gemini 3.5 Live Translate is watermarked with SynthID, an imperceptible watermark embedded directly into the audio. This helps detect AI-generated content to prevent misinformation, supporting responsible AI use.
• •Listening Mode on Android: The Google Translate app on Android introduces a listening mode that streams translated audio directly to the phone's earpiece when held to the ear. This enables private, hands-free translation without headphones, useful in situations like guided tours or quick conversations.

Use Cases

• •Multilingual Meetings: In Google Meet, Gemini 3.5 Live Translate enables conversations across over 2000 language combinations. Participants speaking different languages can communicate seamlessly, with translated speech delivered in near real-time, improving collaboration in global teams.
• •Ride-Hailing Communication: Grab is testing the model for multilingual communication between drivers and travelers during pickups. With over 10 million voice calls monthly, the translation helps bridge language gaps, ensuring smooth and efficient ride coordination.
• •Live Broadcast Dubbing: Broadcasters can use Gemini 3.5 Live Translate to provide live dubbing for international audiences. The model's low latency and natural-sounding output allow viewers to hear translations in their native language with minimal delay, enhancing reach.
• •Education and Lessons: Teachers can deliver live lessons to multilingual students using the model. By translating speech in real time, students hear content in their preferred language, facilitating inclusive education without the need for human translators.
• •Personal Travel Translation: Travelers can use the Google Translate app with headphones to hear translations of spoken language that mirrors the speaker's tone. The listening mode on Android offers private translation via the earpiece, handy in museums or guided tours.

Who is this for?

Developers building voice translation applications using the Gemini Live API, enterprise teams utilizing Google Meet for cross-lingual meetings, global travelers and expatriates needing real-time voice translation, educators conducting multilingual classes, content creators and broadcasters dubbing live content, and service platforms like ride-hailing companies facilitating driver-passenger communication. The model is also aimed at businesses requiring simultaneous interpretation for international conferences and customer support.