据悉，项目已上线35个专属智能体，完成26个原有业务系统的升级改造，在城市公交、应急管理、路桥 养护等场景的应用效果尤为显著。 如在公交领域，公交大模型突破了精准识别"下客"客流的共性难题，识别准确率突破93%，车辆到站时 间预测精准度提升至分钟级，可智能自动调度80%的排班任务，提升运营效率30%以上；在城市应急领 域，AI防汛智能体展现出强大的预警能力，经实测可提前24小时动态预测内涝风险，试点阶段成功预 警14起积水事件，为城市防汛争取宝贵时间；在城市路桥养护领域，智慧巡检系统将实现全覆盖监测、 全方位检查，可精准识别路面病害及绿化缺损问题；在垃圾清运调度领域，实现了清运点自动匹配、垃 圾量合理预测及不安全行为实时识别，准确率超90%，城市生活垃圾清运工作更加高效有序。 日前，经过5个多月的联合攻坚，杭州市城市建设投资集团有限公司(下称"杭州城投")与智谱华章科技 股份有限公司联手打造的杭州城投人工智能产业大模型项目(一期)发布建设成果，多个大模型及智能体 全国首次亮相。 "该大模型以GLM-4.5为核心驱动，创新采用'1+1+3+3'架构，初步构建起功能完备的'AI城投大脑'。"杭 州城投相关负责人介 ...

Core Insights - The collaboration between Hangzhou Urban Construction Investment Group and Zhiyu Huazhang Technology has led to the launch of the Hangzhou Urban Construction AI Industry Model Project, which introduces several innovative AI models aimed at enhancing urban management and services [1][2][6] Group 1: Project Overview - The AI Industry Model Project is structured around a "1+1+3+3" framework, establishing an "AI Urban Construction Brain" that includes a foundational AI operation center and over 30 specialized intelligent agents [2][6] - The project focuses on key business areas such as public transportation, public safety, facility maintenance, and waste management, leveraging AI to improve operational efficiency and service quality [1][2] Group 2: Applications and Benefits - The AI models have significantly improved public transportation services, achieving a 93% accuracy rate in identifying passenger flow and providing minute-level predictions for vehicle arrival times [4][5] - The AI flood prevention agent can predict flooding risks 24 hours in advance by integrating real-time monitoring and meteorological data, successfully issuing alerts for 14 flooding events during trials [5] - The smart inspection system for roads and bridges utilizes AI to identify surface defects, enhancing maintenance efficiency across 491 kilometers of elevated roads and 638 kilometers of urban roads [5][6] Group 3: Strategic Implications - The project represents a transformative shift for Hangzhou Urban Construction, moving from isolated applications to a system-wide integration of AI capabilities, thereby enhancing proactive governance and service delivery [6][7] - Future developments will focus on optimizing models, iterating systems, and deepening application scenarios to create a replicable industry model and intelligent agent matrix [7]