Google's Gemini 3.8 Live Avatar Puts a Lip-Synced Face on the AI
The Verge AI · rss · 2026-09-25
Google's Gemini 3.8 Live update introduces "Live Avatar," letting users converse with an animated AI persona that lip-syncs and changes expressions in real time. It's currently limited to Gemini Enterprise customers.
The avatar supports 97 languages and, per Google, can switch between them "without degrading video fidelity or introducing visual drift." A demo shows it speaking English and Japanese with accurate mouth animation, and it can also surface on-screen information during conversations.
More from Multimodal
- Liquid AI Ships DSpark Draft Model, Speeding Up LFM2.5-VL-3B Decoding Up to 3.13x — JosephJacks_ · 2026-09-25
- Opus 5.5 tops BuildingBench at 86.8, but GPT-6 Astra matches it at 84% lower cost — ZhitingHu · 2026-09-25
- Why video models break physics: trajectory locked in first 10% of denoising — linoy_tsaban · 2026-09-25
- Gemini 3.8 Flash in AI Studio can generate MIDI tracks from text, video or images — DynamicWebPaige · 2026-09-25
- Claude Code directed the most boring movie ever made: full pipeline breakdown for $16.87 — Greedy-Giraffe-4269 · 2026-09-25
- Qwen-Image-2.1 Edit copies faces from reference images instead of blending them — Real-Bed-1815 · 2026-09-25