Google DeepMind has introduced Gemini 3.8 Live with Live Avatar, a feature that brings near real-time visual presence to its native live dialogue models. The announcement follows last week's launch of Gemini 3.8 Live and pairs low-latency streaming video with the company's live speech capabilities.
According to the company, Live Avatar combines near real-time video generation with speech to create an experience that listens, sees, and speaks through a dynamic visual persona. It offers precise lip-syncing, natural expressions, and fluid turn-taking, which DeepMind says can help enterprises expand virtual offerings such as customer service and interactive walkthroughs.
The feature is backed by Gemini's reasoning capabilities through asynchronous tool calling, allowing it to trigger tool calls and fetch data in the background while dialogue continues uninterrupted. DeepMind cites a hotel check-in scenario as an example of handling complex tasks without breaking conversational flow.
Live Avatar also supports native multilingual speech-to-speech synchronization, dynamically adapting lip-sync and expressions and transitioning across 97 languages without degrading video fidelity or introducing visual drift, the company said.
Beyond a library of preset avatars, organizations can generate custom avatars from a high-quality reference image while preserving likeness, brand styling, or character identity. Custom avatar creation is currently available only through enterprise allowlisting. DeepMind said all output is watermarked with SynthID, an imperceptible watermark woven into audio and video to help keep AI-generated content detectable.
Gemini 3.8 Live with Live Avatar is available starting today in Gemini Enterprise, with API documentation offered for developers.