The integration of large models and multimodal foundation models into the low-altitude economy is driving a transformative shift, enabling intelligent, autonomous, and efficient operations for low-altitude vehicles (LAVs). This article provides a comprehensive analysis of the role these large models play within the smart integrated lower airspace system (SILAS), focusing on their applications across the four fundamental networks: facility, information, air route, and service. Our analysis yields several key findings, which pave the way for enhancing the application of large models in the low-altitude economy. By leveraging advanced capabilities in perception, reasoning, and interaction, large models are demonstrated to enhance critical functions such as high-precision remote sensing interpretation, robust meteorological forecasting, reliable visual localization, intelligent path planning, and collaborative multi-agent decision-making. Furthermore, we find that the integration of these models with key enabling technologies, including edge computing, sixth-generation (6G) communication networks, and integrated sensing and communication (ISAC), effectively addresses challenges related to real-time processing, resource constraints, and dynamic operational environments. Significant challenges, including sustainable operation under severe resource limitations, data security, network resilience, and system interoperability, are examined alongside potential solutions. Based on our survey, we discuss future research directions, such as the development of specialized low-altitude models, high-efficiency deployment paradigms, advanced multimodal fusion, and the establishment of trustworthy distributed intelligence frameworks. This survey offers a forward-looking perspective on this rapidly evolving field and underscores the pivotal role of large models in unlocking the full potential of the next-generation low-altitude economy.

Unmanned aerial vehicle (UAV) swarm systems have become a frontier in the field of intelligent unmanned systems. However, the high costs and risks of physical experiments pose significant challenges for professional talent development. Building an efficient training and evaluation simulation platform is key to overcoming this challenge. Existing UAV simulation platforms often rely on traditional input devices such as keyboards and mice, which have limitations in terms of intuitive interaction, immersion, and real-time evaluation capabilities. To address these issues, this paper focuses on the mobile device environment and aims to design and construct a multi-agent system educational platform integrating UAV swarm simulation and human-machine interaction. First, a dual-modal mobile interaction framework based on gestures and voice commands is proposed. Next, a high-precision command recognition system is developed by creating a gesture database, performing data preprocessing, feature fusion, and convolutional neural network (CNN) model training, and integrating a self-trained cloud-based speech recognition model. Finally, a complete platform prototype is developed, enabling users to train in real-time UAV swarm formation control, obstacle avoidance, and other tasks through natural interaction on mobile devices, while simultaneously providing a quantitative evaluation of the operation process. This paper integrates the simulation system, multimodal interaction, and real-time evaluation mechanisms, establishing a unified “trainingevaluation” educational platform, offering an innovative solution for multi-agent system teaching and skill assessment.