Gemini 3.8 Live adds real‑time visual avatar

Google DeepMind announced Gemini 3.8 Live with Live Avatar, extending its live dialogue models with near real‑time video generation. The feature pairs speech with a dynamic visual persona that includes precise lip‑sync, natural expressions and fluid turn‑taking, aiming to make enterprise conversations feel more natural.

Live Avatar processes visual and audio inputs together, supports asynchronous tool calls that continue in the background, and adapts its lip‑sync and expressions across 97 languages without visual drift. Enterprises can choose from preset avatars or generate a custom avatar from a reference image, though custom creation is limited to enterprise allowlisted customers. All output is watermarked with SynthID to keep AI‑generated content detectable. The feature is available today in Gemini Enterprise.