Meituan Enhances 'Xingmou' Large Model: AI Technology Ensures Real-Time Blocking of Back-End Risks
20 hour ago / Read about 0 minute
Author:小编   

Recently, Meituan has undertaken a comprehensive upgrade of its 'Xingmou' vertical-domain multimodal large model, along with its integrated software-hardware service system. Independently developed by Meituan's Core Local Commerce-Intelligent Hardware team, this system was initially launched in April 2025 and made its first public appearance in December of the same year.

The upgraded 'Xingmou' system features a new architecture that effectively tackles the challenges associated with implementing AI technology in food safety scenarios. This advancement facilitates a transformative shift in the platform's risk supervision approach. Currently, the system is undergoing a pilot phase within the platform's food and beverage safety monitoring sector.

So far, 'Xingmou' has carried out a cumulative total of 1.96 billion inspections of the back-end environments of Meituan's food and beverage merchants. It has issued over 2.4 million warnings and prompted the rectification of more than 50,000 issues. In 2026, Meituan intends to expedite the deployment of the 'Xingmou' service system across its core business scenarios. The research and development team will concentrate on resolving complex challenges within the food safety domain, aiming to further bolster the capabilities of both the model and the service system.