Google DeepMindTuesday · June 9, 2026FREE

Fluid, natural voice translation with Gemini 3.5 Live Translate

Google DeepMind announced Gemini 3.5 Live Translate, a new voice translation capability built into the Gemini 3.5 model. Unlike traditional translation systems that process speech in discrete chunks, Live Translate handles audio streams continuously, preserving natural prosody, emotion, and speaker identity. The system can translate between multiple language pairs in real-time, with latency low enough for natural conversation flow. DeepMind emphasizes that the translation maintains the speaker's tone and intent, avoiding the flat, robotic output common in previous systems. For developers, this opens up possibilities for more human-like multilingual interactions in applications such as live customer support, international meetings, and content localization. The model is available through the Gemini API, with pricing details to be announced. This release builds on Google's broader push to integrate multimodal capabilities into its AI offerings, following the earlier launch of Gemini 3.5 with improved reasoning and context handling.