Google introduces Gemini 3.1 Flash Live, their highest quality audio & voice model to date. Through it, conversations with Gemini Live and Search Live will be more realistic and faster. It is now more accurate in identifying the tone and tempo of the user's voice and itself speaks like a real human.
The latency before the answer is given is also lower than Gemini 2.5 Flash Native Audio. In addition, improvements have been made where it can distinguish the user's voice from environmental noise such as traffic and TV in the background.
Gemini 3.1 Flash Live supports over 90 languages for real-time multimodal conversations including Bahasa Melayu. This model is accessible to Gemini Live and Search Live users in over 200 countries or regions globally today.
