Ant Group’s Lingbo Tech launched LingBot-Depth 2.0, a spatial perception model trained on 150M samples, achieving 12 first places in depth completion benchmarks and halving depth error in challenging indoor scenes. It also released LingBot-Vision, a visual foundation model with boundary-structure pre-training. LingBot-Depth 2.0 passed Orbbec Depth Vision Lab certification, with integrated camera products expected by year-end.

Comments (0)