At a Glance · AI-generated

Google launched Gemini 3.1 Flash Live on March 26, a real-time multimodal voice model designed to enable more natural-sounding audio interactions with artificial intelligence. The new model supports low-latency audio and video processing alongside tool use for AI agents, marking a significant advancement in conversational AI capabilities. Google characterized the release as its most natural-sounding AI voice model to date, though the technology's improved realism raised questions about the ability to distinguish human conversation from machine interaction.