SANA-WM is worth watching for one reason: it combines longer video generation with explicit camera control.
Five quick facts:
- It is an open-source 2.6B-parameter world model from NVIDIA Research.
- It targets minute-scale 720p video generation.
- It uses precise 6-DoF camera trajectories instead of only unconstrained motion.
- The paper reports single-GPU generation, with a distilled variant denoising a 60-second clip in about 34 seconds on one RTX 5090.
- Its benchmark reports roughly 36x higher throughput than prior open-source baselines.
If you want the official demos, paper, and code in one place, I collected the primary resources here:
Official references:
- Project page: https://nvlabs.github.io/Sana/WM/
- Paper: https://arxiv.org/abs/2605.15178
- Code: https://github.com/NVlabs/Sana
Top comments (0)