The recent advancements in the application of large language models (LLMs) within the medical domain have shown significant promise, particularly in specialized tasks such as Traditional Chinese Medicine (TCM), pharmaceutical patent analysis, and medical text processing. The integration of prompt engineering frameworks, such as TCM-Prompt, has demonstrated enhanced performance in complex medical tasks, including disease classification and herbal medicine recommendation. Additionally, the development of specialized LLMs like MedGo has set new benchmarks in Chinese medical information processing, achieving top performance in public evaluations and practical deployment in clinical settings. The introduction of intelligent agents like PatentAgent for automated pharmaceutical patent analysis marks a transformative step, offering comprehensive solutions from patent reading to core chemical identification. Furthermore, the utilization of LLMs in medical information extraction and query generation for clinical trials has shown comparable or superior effectiveness to expert-created queries, emphasizing the potential for efficient and accurate medical decision support. These developments collectively underscore the transformative impact of LLMs in advancing precision and efficiency in medical research and practice.