Google Adds Live Avatar to Gemini 3.8 Live, an Animated Video Layer for Real-Time Conversation

Google launched Live Avatar, a visual layer for Gemini 3.8 Live that pairs near real-time video generation with the model’s speech capabilities, giving it an animated on-screen presence during conversations.

Details

• What it does: creates an animated character with precise lip-syncing and natural expressions that converses in near real time, processing visual and audio input simultaneously
• Background task handling: supports asynchronous tool execution, so the avatar can keep talking naturally while fetching data in the background — Google’s example is handling a hotel check-in without breaking the conversation
• Language support: works across 97 languages with seamless mid-conversation switching and lip-sync that adapts to the language being spoken
• Customization: avatars can be generated from reference images while preserving brand identity, alongside several preset avatars with distinct appearances and voices
• Trust and safety: all Live Avatar output is watermarked with SynthID to keep AI-generated video identifiable
• Availability: launched September 24, 2026, exclusively through Gemini Enterprise, with custom avatar creation requiring enterprise allowlisting; no pricing disclosed

What happened next

Live Avatar extends Gemini 3.8 Live — launched just over a week earlier — with a visual front end aimed at enterprise use cases like customer service and virtual assistants, following the same pattern as Gemini 3.8’s earlier text-to-speech launch of pairing new audio capability with an enterprise-first rollout before wider availability.