Ant Group’s robotics unit Robbyant has released model weights and inference code for LingBot-Touch, a model designed to help robot hands use touch to handle objects. It was trained on nearly 5,000 hours of recordings from three types of tactile sensors so that it could work across different hardware.
Robbyant evaluated the model on 26 real-world tasks using two robotic hands. The tests included handling fragile objects and coordinating both hands. [Robbyant]
source & further reading
technode.com — original article
Tencent expands Hy4 inference capacity tenfold as WorkBuddy demand surges
DeepSeek plans official plugin marketplace as Harness daily users reach 1.54 million
Ant launches Ling-3.2-flash multimodal model with million-token context