The integration of biomedical big data with artificial intelligence (AI) technologies represents a pivotal approach to constructing a comprehensive and adaptive training system for medical students. By developing an AI-empowered integrated teaching platform encompassing five core modules - "structured medical record automatic generation, competency profiling, knowledge-based question answering, intelligent assessment, and decision-making" - we enrolled and trained 200 students who rotated through the department of neurology, the Second Affiliated Hospital of Anhui Medical University, while verifying the platform's application efficacy. Results demonstrated that students in the observation group achieved significantly higher scores than those in the control group across dimensions such as perceived ease of use, perceived usefulness, and learning engagement. Additionally, favorable outcomes were observed in terms of optimal path selection rate and average time consumption in decision-making simulations, performance in knowledge-based interactive question answering, application accuracy, and response speed. This novel teaching model enhanced students' comprehensive clinical competencies, refined their diagnostic and therapeutic thinking, and strengthened their ability to conduct standardized diagnosis and treatment. Concurrently, students' understanding of the breadth and depth of medicine was continuously deepened, ultimately yielding desirable educational outcomes. Therefore, the AI large-model-empowered collaborative teaching system based on biomedical big data can provide a new paradigm for realizing precision and adaptive medical education.