Introducing Gemini 3.8 Live with Live Avatar
Software engineers are thrilled to announce the launch of Gemini 3.8 Live with Live Avatar, a cutting-edge feature that brings near real-time visual presence to live dialogue models. This innovative tool combines speech and video generation, resulting in a dynamic visual persona that listens, sees, and speaks simultaneously.
The Live Avatar feature enhances enterprise offerings by providing engaging customer service and interactive walkthroughs, making digital exchanges more interactive and accessible. With precise lip-syncing, natural expressions, and fluid turn-taking, Live Avatar elevates the conversational experience to new heights.
Gemini 3.8 Live with Live Avatar supports a wide range of characters, each with a unique look, voice, and expressive presence. Conversations are inherently multimodal, requiring the ability to listen, look, speak, and use facial expressions. By processing visual and audio inputs simultaneously, Live Avatar generates enriching conversations that feel more natural and comprehensive.
Underlying the Live Avatar feature is Gemini's advanced reasoning. The feature utilizes asynchronous tool calling, enabling it to trigger tool calls and fetch data in the background while maintaining an uninterrupted dialogue. This allows Live Avatar to handle complex tasks, such as checking in a hotel guest, without disrupting the conversation flow.
Live Avatar is designed to be language-independent, featuring native multilingual speech-to-speech synchronization. The feature dynamically adapts its lip-sync and expressions, seamlessly transitioning across 97 languages without compromising video fidelity or introducing visual drift. This ensures that organizations can communicate effectively in their preferred language, without any loss of quality or visual coherence.
Organizations can further personalize their Live Avatar experience by customizing it to fit their brand identity. Developers can generate a fully animated, responsive avatar using a high-quality reference image, preserving the reference likeness, brand styling, or character identity. Currently, custom avatar creation is available through enterprise allowlisting.
To ensure responsible deployment of AI-generated content, Live Avatar is equipped with strict safeguards, such as the SynthID watermark. This imperceptible watermark is woven directly into the audio and video output, helping to minimize misinformation and misattribution. To learn more about Gemini 3.8 Live with Live Avatar's comprehensive approach to safety and responsible deployment, refer to the model card.
Written by urgent.news from Google DeepMind's reporting — not their text. Machine-written — may contain errors; check the original before relying on it.
This story
This is one outlet's version. Read the fullest account.
- Introducing Gemini 3.8 Live with Live Avatar blog.google