Google launches Gemini 3.8 Live with real-time animated avatars

Google has announced the general availability of ‘Gemini 3.8 Live with Live Avatar’ for Gemini Enterprise customers. This update integrates the Gemini 3.8 Live conversational model with real-time video generation, allowing the AI to present an animated digital persona that features precise lip-syncing and natural facial expressions.

The technology supports 97 languages and can switch between them without losing video quality or visual synchronization. The avatars can process visual inputs, such as camera feeds or screen sharing, and perform asynchronous tool calls in the background—enabling the AI to complete tasks like hotel check-ins while maintaining a continuous conversation.

Businesses can choose from a library of preset avatars or create custom ones using a single high-quality reference image and voice sample. To prevent misuse and impersonation, custom avatar creation is restricted to an approved list of companies, and all generated content is embedded with ‘SynthID’ digital watermarks for identification. The service is currently available via the Google Cloud console and Live API for endpoints in the US and EU.