Gemini 3.1 Flash Live: Making audio AI more natural and reliable
Google introduced Gemini 3.1 Flash Live, a high-quality audio model offering improved precision and lower latency for natural, reliable real-time dialogue.Summary
Google has launched Gemini 3.1 Flash Live, its highest-quality audio and voice model designed for next-generation voice-first AI interactions, emphasizing speed and natural rhythm. This model is available to developers via the Gemini Live API in Google AI Studio, to enterprises through Gemini Enterprise for Customer Experience, and to the general public via Search Live and Gemini Live. For developers, 3.1 Flash Live shows improved reliability in complex task execution, scoring highly on benchmarks like ComplexFuncBench Audio and Scale AI’s Audio MultiChallenge, and features better tonal understanding to handle nuances like pitch and pace dynamically. For everyday users, Gemini Live now offers faster, more natural responses and can maintain conversational context twice as long. Furthermore, the model's inherent multilingual capabilities support the global expansion of Search Live to over 200 countries. Crucially, all audio generated by 3.1 Flash Live is watermarked using SynthID to reliably detect AI-generated content and help prevent misinformation.
(Source：Gemini)