SANA-Video 2.0

Other: deep learning(nvlabs.github.io)view on HackerNews
SANA-Videovideo generationdeep learninghybrid attentionattention residualsefficient video generationNVIDIA ResearchAImachine learning

Author: ilreb

Date: 7/24/2026

Article Summary:
This article presents SANA-Video 2.0, a hybrid video diffusion transformer that generates high-quality video up to 720p on a single GPU, achieving a VBench score of 84.30 in 13.2s at 480p and outperforming larger softmax video DiTs at a fraction of the latency.