ainewsblitz.com

Breaking

Alibaba-Quark Open-Sources Live Avatar, a 14B Diffusion Model for Real-Time, Infinite-Length Talking Avatars

Alibaba-Quark Open-Sources Live Avatar, a 14B Diffusion Model for Real-Time, Infinite-Length Talking Avatars
  • Media Generation
  • Open Source
  • Foundation Models

Alibaba-Quark has released Live Avatar, a 14-billion-parameter diffusion model that generates audio-driven talking avatars in real time and, crucially, for effectively unlimited durations. The project, accompanied by an arXiv paper and a demo-rich project page, was accepted as an Oral presentation at ECCV 2026 and is now available under the permissive Apache 2.0 license, with model weights on Hugging Face and inference code on GitHub.

Continue reading

The rest of this article is for AI News Blitz readers. Choose an option below to keep reading.

$20
Read this article
$29/month
Unlimited — all 27,863 articles, the full archive, and comprehension quizzes
Save 72%
$98/year
≈ $8.17/month
Unlimited, billed once a year