Gemini 3.8 Live Adds Live Avatar for Enhanced Enterprise Conversations

What changed

Google DeepMind has launched Gemini 3.8 Live with Live Avatar, an extension of its Gemini 3.8 Live conversational AI. This new feature introduces near real-time visual presence, allowing Gemini to respond with a dynamic visual persona that includes speech and video generation. The Live Avatar aims to create a more natural and intuitive conversational experience by pairing near real-time video generation with speech, enabling precise lip-syncing, natural expressions, and fluid turn-taking. This capability is now available within Gemini Enterprise.

The feature enhances multimodal conversations by processing visual and audio inputs simultaneously, generating richer dialogue. It supports asynchronous tool execution, allowing the AI to trigger tool calls and fetch data in the background while maintaining an active dialogue, ensuring an uninterrupted conversational flow for complex tasks. Furthermore, Live Avatar offers native multilingual speech-to-speech synchronization, adapting lip-sync and expressions seamlessly across 97 languages without degrading video fidelity. Organizations can also customize avatars using a high-quality reference image to match brand identity, though custom avatar creation is currently available via enterprise allowlisting.

Trust and transparency are addressed through watermarking all AI-generated output with SynthID, an imperceptible watermark embedded in the audio and video to ensure detectability and minimize misinformation. A model card detailing the approach to safety and responsible deployment is also available.

Why it matters for builders

For AI builders, Gemini 3.8 Live with Live Avatar provides tools to create more immersive and interactive enterprise applications. The integration of visual presence, alongside advanced reasoning and asynchronous tool handling, allows for the development of sophisticated agents capable of managing complex tasks while maintaining engaging user interactions. The extensive multilingual support and customizable avatar options offer flexibility for global deployments and brand alignment.

Practical impact

Developers can leverage Gemini 3.8 Live with Live Avatar within Gemini Enterprise to build next-generation customer service bots, interactive walkthroughs, and virtual assistants. The ability to generate responsive avatars that match brand aesthetics and communicate fluently in multiple languages can significantly enhance user engagement and accessibility. Builders should explore the API documentation to integrate these advanced visual and conversational capabilities into their applications, focusing on use cases that benefit from a more human-like AI presence.

Caveats and source limits

The source material indicates that custom avatar creation is currently available only through enterprise allowlisting. While the feature supports 97 languages, specific details on performance benchmarks or independent evaluations of its real-time capabilities and lip-sync accuracy across all supported languages are not provided. The exact pricing for Gemini Enterprise, which includes this feature, is also not detailed in the announcement.