Meituan releases LongCat-Video-Avatar 1.6 for stable long-form audio-driven video generation

·2026.11.15 22:41

Key point

The update improves lip-audio synchronization and reduces error accumulation to support coherent video generation over 10 minutes.

Meituan released LongCat-Video-Avatar 1.6, an open-source framework upgrade for audio-driven human video generation.
The model supports stable video generation exceeding 10 minutes by mitigating error accumulation during autoregressive continuation.
Improvements include more natural lip animation, reduced temporal instability, and better coordinated facial and body motion.
End-to-end pipeline and model execution optimizations enable faster inference and more efficient deployment.

This summary was generated automatically by AI. Check the original for the author's claims and context. Copyright belongs to the original author.

Our guide explains how the AI works. Report summary errors, attribution issues, or removal requests via Contact.