Gemini 3.8 Live introduces real-time visual avatars for enterprise applications

Google has announced Gemini 3.8 Live with Live Avatar, adding real-time video generation and speech to its conversational AI models.

2 min read

Collage showcasing various Gemini 3.8 Live avatars and user interfaces.
Gemini 3.8 Live introduces real-time visual avatars for enterprise applications.

Google has introduced Gemini 3.8 Live with Live Avatar, a new feature that brings near real-time visual presence to conversational artificial intelligence. The system natively couples live dialogue capabilities with low-latency streaming video, creating a visual persona that listens, sees, and speaks with natural expressions and precise lip-syncing.

Key facts

Model Name
Gemini 3.8 Live
Languages Supported
97 languages
Endpoints
US and EU endpoints
Audio Model Alternatives
Gemini 3.5 Transcribe
Model: Specialized, modular audio capabilities
Specialization: —
Gemini 3.5 Live Translate
Model: Specialized, modular audio capabilities
Specialization: —
Gemini 3.8 Flash-Lite TTS
Model: Specialized, modular audio capabilities
Specialization: —
Gemini 3.8 Flash TTS
Model: Specialized, modular audio capabilities
Specialization: —
UI screenshot of Gemini 3.8 Live handling insurance claim notes in real time.
Insurance claim live agent team interface showing background tool execution notes., Source ↗

Multimodal conversation

Human conversation relies on multiple senses at once, including listening, looking, speaking, and observing facial expressions. The Live Avatar feature processes visual and audio inputs simultaneously to support more comprehensive and intuitive interactions. Enterprises can use the tool for customer service or interactive walkthroughs.

Product screenshot showing the live transcript and video interaction for Gemini 3.8 Live.
Live transcript and video interaction interface demonstrating multimodal conversation., Source ↗

Asynchronous tools

The feature is backed by advanced reasoning capabilities that support asynchronous tool execution. This allows the model to trigger tool calls and fetch data in the background while continuing an active dialogue, ensuring an uninterrupted conversational flow during complex tasks such as hotel guest check-ins.

Global scale

To support international users, the system features native multilingual speech-to-speech synchronization. It dynamically adapts its lip movements and expressions to seamlessly transition across 97 languages without degrading video fidelity or introducing visual drift.

Customization and safety

Organizations can choose from a library of preset avatars or create custom avatars using a high-quality reference image to match specific brand identities and styling. Custom avatar creation is currently available through enterprise allowlisting.

To maintain transparency, all audio and video outputs are marked with SynthID, an imperceptible watermark woven directly into the media to help detect AI-generated content and minimize misinformation. Gemini 3.8 Live with Live Avatar is available starting today in Gemini Enterprise.

Source: Original announcement

Worth a read?

Report

Comments