Natura-1
$ 0
Natura-1 is Nature Will’s first general robot foundation model. It learns from huge amounts of human video and turns everyday instructions into skilled two-handed actions.
Available on backorder
Brain Score
4
Specifications and details:
| Nationality | China |
|---|---|
| Website | https://www.naturewill.com/ |
| Model type | Foundation Model |
| Manufacturer | Nature Will |
| Release date | October 13, 2026 |
Description
Natura-1 is the first foundation model from Nature Will, a physical AI lab based in Beijing, China. The company calls it a general model for the physical world, and it shared the model as a research preview in October 2026. Natura-1 learns mostly from human experience. For example, its training draws on more than one million hours of first-person video of people doing everyday tasks. It then adds only a small amount of robot data on top. This approach matters because robot data is slow and costly to collect, while human video is plentiful.
Natura-1 lets a robot follow spoken or written instructions and act on them with two dexterous hands. It can also handle tasks it has never seen, and it adjusts and retries when something goes wrong. In company demos, the robot makes wontons, packs boxes and serves tea over several minutes. The model also predicts what the scene will look like next, so it works partly as a world model. However, Nature Will has not released the weights, an API or a sign-up path yet. Instead, the preview marks the lab’s first public step toward general physical intelligence.
New! 2026 Humanoid
Robot Market Report
198 pages of exclusive insight from global robotics experts — uncover funding trends, technology challenges, leading manufacturers, supply chain shifts, and surveys and forecasts on future humanoid applications.
now Google DeepMind
Contact Humanoid.guide
Website: https://www.naturewill.ai/blog/natura-1?lang=en
