Lower latency and improved precision define the latest update to Gemini 3.1 Flash Live. Google DeepMind tuned the voice model to make interactions feel more fluid. This incremental update targets reliability over new capabilities. Developers can now deploy voice interfaces with reduced lag and higher response accuracy for real-time users.