Google Launches Gemini 3.5 Live Translate for Real-Time Natural Dialogues

Google has introduced the latest version of its AI speech translation model, Gemini 3.5 Live Translate, designed to enable natural multilingual conversations in real time. Unlike existing solutions that process speech sequentially, the new model operates in a continuous flow, drastically reducing latency and making dialogues smoother.

The model automatically detects languages and supports over 70, making it versatile for various scenarios—from business negotiations to travel interactions. Gemini 3.5 Live Translate is optimized for noisy environments, overlapping voices, and informal speech, positioning it as a universal tool for both professional and everyday use.

Developers have prioritized high-quality speech synthesis, ensuring the system not only translates text but also preserves the original speaker’s intonation, pace, and emotional tone. This approach makes translated speech sound more natural and comprehensible. The technology is now accessible to partners and developers for integration into diverse applications and services.

The launch of Gemini 3.5 Live Translate aligns with the growing trend of AI-powered translators becoming integral to daily life. Competitors like Apple (AirPods Pro 3 with Live Translation) and T-Mobile (AI call translation testing) have also introduced similar solutions. Google is further expanding its Google Translate service by adding support for translation via headphones on Apple and Android devices.