Dyna Robotics launches Dyna-2: General-purpose physical agents with hour-long autonomy
Current overview (not historical)
Dyna Robotics has introduced Dyna-2, a general-purpose physical agent featuring a world action model trained on one million hours of human video. The system enables robots to learn new tasks from just 45 minutes of demonstration and execute long-horizon folding tasks lasting up to 24 hours in real-world settings like restaurants. Dyna-2.1, the accompanying physical robot agent, utilizes a half-humanoid wheeled base for enhanced stability and has been deployed in venues such as Din Tai Fung, where it autonomously folds thousands of napkins daily. The robot achieved a 99.4% success rate in single-folding tasks.
Written with AI from the reporting · updated 10/23/2026
Expand the detailed fact route and sources
Story route
A station-by-station route from the first public evidence onward; every station carries its own evidence and uncertainty.
- 1Unclassified fact stationOct 14, 00:00recorded Oct 14, 22:22single sourcerevised
Dyna-2: General Purpose Physical Agents with Hour-Long Auton
Read original source ↗ Latent SpaceReplay this moment in the explorerEvidence & revisions · 1 source records
- Subject：
- Dyna Robotics
- Action：
- 发布 Dyna-2 世界行动模型及 Dyna-2.1 物理机器人代理
- Object：
- world-action model 预测未来视频以指导机器人动作；实现 24 小时连续折叠餐巾 99.4% 成功率；从 45 分钟演示即可学习新任务
- Occurred (fact side)：
- Oct 14, 00:00
- Published (valid side)：
- Oct 14, 06:25
- First seen (known side)：
- Oct 14, 22:22
Fact stationSingle source · one-sided claimRevisedCopy may have been revised later; time replay is bounded. Layer labels group reading only, not causation between layers.
Uncertainty marks flag where to check the sources; they are not proven causation.
Report timeline
Explore every side through dated source reports.
- Latent SpaceRelated fact: single sourceSource report revisedDyna Robotics deploys Dyna-2.1 physical agent with hour-long autonomy based on million-hour video model
Report first recorded Oct 14, 22:22
Dyna Robotics released the Dyna-2.1 physical robot agent, built upon a world-action model pre-trained on over one million hours of human video. The robot can learn new tasks from as little as 45 minutes of demonstrations and demonstrated 24 hours of continuous napkin folding in a restaurant setting, including recovery from errors. Dyna Robotics deploys these robots, which use wheeled semi-humanoids, for daily tasks like folding thousands of napkins at Din Tai Fung.
Read original source ↗
Event discussion
Reader opinions are hosted on GitHub Discussions; posting requires a GitHub account. Comments are not verified facts.