LongCat-Video-Avatar 1.5 Officially Open-Sourced
Original title:从高拟真到真可用,LongCat-Video-Avatar 1.5 正式开源
AI summary
Meituan officially open-sourced LongCat-Video-Avatar 1.5, achieving comprehensive leaps in lip sync, physical plausibility, long-video stability, multi-person interaction and efficient inference. The model uses DMD distillation to compress generation steps to 8, improving inference efficiency by about 15 times and generating a 10-second video in about 1 minute; Wav2Vec2 is upgraded to Whisper-large, and GRPO frame-by-frame preference alignment is introduced to optimize hand stability.
Key Takeaways
- Meituan officially open-sourced LongCat-Video-Avatar 1.
- 5,achieving comprehensive leaps in lip sync,physical plausibility,long-video…
- The model uses DMD distillation to compress generation steps to 8,improving inference efficiency…
- Wav2Vec2 is upgraded to Whisper-large,and GRPO frame-by-frame preference alignment is introduced…
Quick Facts
- Published
- Updated
- Source
- 公众号:龙猫LongCat(美团)
- Category
- Model Releases
- AI score
- 88
开源数字人模型首次在多人场景区分、长视频稳定性和推理效率上同时达到商业产品水平,评测透明度高。
Source & copyright
This article is based on curated titles and AI summaries from the AIHOT public API, with attribution and links back. Original copyright remains with the original publishers and authors; this site does not mirror full texts.
How to Cite
When citing this article, keep the title, date, and permalink, and credit Orange News as the source.
Orange News: LongCat-Video-Avatar 1.5 Officially Open-Sourced. May 22, 2026. https://new.nat6.net/en/news/cmsdu2t2b011wrofzn9v6zcye/