Google continues to refine Gemini, focusing especially on improving voice interaction to achieve more natural and useful conversations, even with complex questions. In October, Gemini Live received a significant update, and now Google is working on new features that aim to bring the experience even closer to a real conversation. Gemini Live currently operates with the Gemini 2.5 Flash model, known for its speed and fluency in responses. However, with the arrival of Gemini 3 Pro in November, the company introduced "Labs" features, allowing users to test experimental capabilities before their general release. According to 9to5Google, version 17.2 of the Google app reveals that Gemini Live will be one of the main beneficiaries of these updates. Among the most notable improvements is the "Live Thinking Mode," a feature that gives Gemini Live more time to process and formulate voice responses. Thus, instead of responding immediately, the AI will be able to provide more complete and detailed explanations, especially useful for complex questions. This mode would leverage the more advanced models of Gemini 3, such as Pro and those specialized in reasoning. Additionally, Google is testing new experimental features for Gemini Live, many of which already exist in the Gemini 3 Flash and Pro chat, such as the use of conversation history and integration with Personal Intelligence applications. The novelty is that these capabilities will also come to the voice experience. Finally, Labs is exploring options like allowing an agent to control the mobile device to perform tasks or delegate complex research, features still in the testing phase. All of this suggests that Gemini Live is about to make a significant leap on Android, establishing itself as an increasingly advanced and versatile tool. Source: 20minutos.es