Ant Lingbo Open-Sources LingBot-Touch Tactile Foundation Model for Robots

On November 10, Ant Lingbo Technology open-sourced LingBot-Touch, a tactile foundation model intended to supply robot “brains” with a missing layer of touch perception. The company says the model was pretrained on nearly 5,000 hours of real tactile data.

LingBot-Touch covers three mainstream sensing principles: magnetic tactile, vision-based tactile, and piezoresistive sensing. It converts signals from these different hardware types into a single tactile representation, allowing the same model output to be used by various dexterous hands and action models.

The release is aimed at reducing fragmentation in tactile learning, where data collected on one hand or sensor often does not transfer to another. By mapping heterogeneous touch signals into a shared space, LingBot-Touch seeks to let developers reuse tactile knowledge across robot embodiments and downstream action policies.

According to Ant Lingbo, this unified representation helps robots complement clear vision and capable planning with the tactile feedback needed for fine manipulation. The open-source release is positioned as infrastructure for reuse across hardware, models, and tasks, rather than a single-robot solution.