DeepMind has introduced Gemini 3.1 Flash Live, a voice model designed to make spoken interactions with AI feel more natural. According to the company, the model delivers improved precision and lower latency, which should help reduce awkward pauses and misunderstandings during live conversation.
DeepMind describes the intended effect as voice AI that is more fluid, natural, and precise. The emphasis on reliability suggests the model is aimed at applications where real-time audio matters, such as voice assistants or live translation.
As with any vendor announcement, these claims have not been independently verified. Still, the update signals that DeepMind is continuing to focus on the audio side of AI, not just text and image generation.