Science & Tech · AI · 58 mins ago
Ant Group’s Robbyant releases tactile AI model for robot hands
Ant Group’s robotics unit Robbyant released LingBot-Touch, an AI model designed for robot hands.
The model uses touch signals to help hands handle objects, rather than relying only on sight.
Robbyant trained it on nearly 5,000 hours of recordings from three types of tactile sensors.
The aim was to make the model work across different sensor hardware.
In tests, it completed 26 real-world tasks using two robotic hands, including handling fragile objects and coordinating both hands.
Releasing the model weights and inference code lets others use the system, though no further plans were stated.
Ant Group robotics unit Robbyant released LingBot-Touch, a model designed to help robot hands handle objects using touch.
The company made the model weights and inference code available as open source.
LingBot-Touch was trained on nearly 5,000 hours of recordings from three types of tactile sensors.
Robbyant evaluated the model on 26 real-world tasks using two robotic hands.
The tests included handling fragile objects and coordinating both hands.
- Who
- Ant Group’s robotics unit Robbyant.
- What
- It released LingBot-Touch, an open-source tactile AI model for robot hands.
- When
- 2026-11-10.
- Where
- Not stated.
- Why
- The model is designed to help robot hands use touch to handle objects.
This story does not have two clearly opposing sides.
This story does not have a timeline yet.
- Model
- LingBot-Touch
- Developer
- Robbyant, Ant Group’s robotics unit
- Training data
- Nearly 5,000 hours of recordings
- Sensor types
- Three
- Evaluation
- 26 real-world tasks using two robotic hands