Lower latency and improved precision define the latest update to Gemini 3.1 Flash Live. Google DeepMind optimized the voice model to reduce lag during real-time interactions. This incremental update makes audio responses feel more fluid. Developers can now build voice interfaces that respond with higher reliability and less audible hesitation.