Gemini 3.8 Live (ai.google.dev)

Gemini 3.8 Live has been launched as the new default model for low-latency voice interactions, enhancing real-time dialogue capabilities without delays. This version introduces several significant features, including interleaved reasoning and asynchronous function calling, which allow for a more fluid user experience. The model supports various input types—text, images, audio, and video—with impressive token limits of 131,072 for inputs and 65,536 for outputs, catering to complex interaction needs. Proactive audio is now a standard feature, eliminating the need for manual configuration, while support for affective dialogue has been removed. This update is vital for the AI/ML community as it sets a new benchmark for voice agent capabilities and asynchronous workflows, making it easier for developers to create interactive applications. The move to asynchronous execution as a default mode increases efficiency in processing, allowing developers to leverage features like client content updates and audio streaming throughout the session. These updates not only enhance user engagement but also pave the way for innovative applications in voice technology, further solidifying Gemini's position in the competitive landscape of AI-driven communication tools.