Introducing Gemini 3.8 Live with Live Avatar
Google announced Gemini 3.8 Live with Live Avatar on September 24, 2026, pairing its live speech models with near-real-time video so the AI can appear as a visual persona. It is available to enterprises via Gemini Enterprise, supports speech in 97 languages and includes SynthID watermarking; custom avatars are limited to approved customers.
What the article says
On September 24, 2026, Google introduced Gemini 3.8 Live with Live Avatar, building on the launch of Gemini 3.8 Live the previous week. The addition brings near real-time visual presence: video generation is paired with speech so an avatar can listen, see and respond.
The post cites precise lip-syncing, natural expressions, simultaneous processing of visual and audio input, and asynchronous tool calls that run in the background during the conversation, for example to check in a hotel guest. Lip-sync and expressions adapt across 97 languages. A library of preset avatars is offered; custom avatars from a reference image are limited to allowlisted enterprises. Outputs are watermarked with SynthID. The feature launched that day in Gemini Enterprise, with API documentation for developers. Suggested uses include customer service and interactive walkthroughs.
Read the official article AnnuIA profile
Summary written by AnnuIA in its own words from the official page of Google DeepMind. The content and rights belong to the publisher: please read the original article.