Ant Group’s Robbyant Open-Sources LingBot-Vision: A 1B Boundary-Centric Vision Foundation Model for Dense Spatial Perception

Ant Group's research team, Robbyant, has released LingBot-Vision, a one-billion-parameter computer vision model trained using masked boundary modeling. By prioritizing spatial edges as a core training signal, the model achieves performance levels comparable to significantly larger architectures in dense visual perception tasks. This open-source release serves as the foundational backbone for the company’s LingBot-Depth 2.0 system.