SANA-Video 2.0 pairs hybrid attention with 84.30 VBench and major speedups
danijarh · x · 2026-07-24
SANA-Video 2.0 is a newly released video model optimized end-to-end for efficiency while keeping quality high.
It uses a hybrid linear-softmax attention design, Block Attention Residuals, and the Sol-Engine acceleration stack. The team says it trained unified 5B and 14B models from scratch on limited resources — 16 H100 nodes for the 5B model and 48 B200 nodes for the 14B model — and reports 84.30 VBench total, 3.2× faster DiT forward passes at 720p/60s, 13.06s for 720p/5s on a single H100, and 120× faster generation than Wan 2.2-A14B under the same setup.
Related event: NVIDIA Unveils SANA-Video 2.0 for Efficient Video Generation(2 posts)→
More from Multimodal
- Open-source local studio bundles image, video, music and AI movie maker offline — lelus78 · 2026-10-26
- Techhalla shares full 2D-to-3D workflow with Opus 5.5 and Kling 4.0 prompts — techhalla · 2026-10-26
- Rebuilding real events as photoreal SF digital twins, filmed cinematically from a phone — bilawalsidhu · 2026-10-26
- How to turn Claude Code into a video editor: a full reel in 19 minutes — PrajwalTomar_ · 2026-10-26
- "Recursive eyes" prompt idea produces mesmerizing spiral of infinitely recurring faces — LudovicCreator · 2026-10-26
- Black Forest Labs open-sources 5.6B FLUX 3 Image Klein, +91 Elo with 38% fewer params — dl_weekly · 2026-10-26