AI Summary

XPENG and researchers from Peking University, The University of Hong Kong and other academic partners have released GroundingPI, an open‑source visual perception foundation model. The model, built with roughly 4 billion parameters, is designed to let AI systems accurately identify objects and determine their spatial positions in the physical world. In testing, GroundingPI attained the highest aggregate score among 44 baseline models across 34 perception benchmarks covering 11 capability categories. The effort reflects XPENG’s support for advancing embodied intelligence through open research.

AI summaries can be wrong sometimes—always verify important details using the source article.