Google adds talking video avatar to Gemini 3.8 Live

On September 24, 2026, Google made Live Avatar generally available — an add-on to Gemini 3.8 Live that generates video with a talking character in near real time, syncing lips and facial expressions across 97 languages.

Compiled by Sergey KostenchukPublished 2026-09-25Updated 2026-09-25

2026-09-25 Multimodal Google

🎨 Google adds talking video avatar to Gemini 3.8 Live

Live Avatar is an add-on to the Gemini 3.8 Live voice model: video with a talking character is generated in near real time, with synced lips and facial expressions adapted for 97 languages. The feature reached GA in Gemini Enterprise and via the Gemini Live API.

🌍 Video avatars are moving from demo to production: Cox Automotive has integrated them into Autotrader, Equal AI handles over a million calls a day, and Salesforce is connecting them to Agentforce.

👤 A developer can build an agent that sees, hears, talks with a face, and calls your APIs in the background. One minute of avatar speech costs about 39 cents, with billing only for speaking; all output is watermarked with SynthID.

Source 1: https://blog.google/innovation-and-ai/models-and-research/gemini-models/gemini-3-8-live-with-live-avatar/ Source 2: https://cloud.google.com/blog/products/ai-machine-learning/gemini-3-8-live-with-live-avatar-is-now-generally-available

Sources