"Beijing Medical Thousand Inquiries" medical large model fully open source, comprehensive score ranked first in the MedBench evaluation list.

Recently, JD HEALTH (06618) launched the "Jing Yi Qian Xun" medical large-scale model, becoming the first comprehensive open-source vertical large-scale model in the domestic medical industry. At the same time, in the updated MedBench evaluation list on February 24th, the "Jing Yi Qian Xun" medical large-scale model ranked first with a comprehensive score of 96.1; it also scored 88.9 on the authoritative medical evaluation set, MedQA. The comprehensive open-source launch means that all model parameters, data structures, multi-stage model training, model evaluation, and other core code of the "Jing Yi Qian Xun" medical large-scale model will be made public, demonstrating JD HEALTH's determination in promoting the advancement of medical AI technology and driving the transformation and development of the medical industry through medical AI. As of now, the "Jing Yi Qian Xun" medical large-scale model has become the most widely used vertical large-scale model with the richest application scenarios, the deepest collaboration with hospitals, the most cooperative doctors, and the widest participation of users in the industry. It has pioneered the large-scale application of models in the medical and health industry, achieving large-scale and full-scenario landing applications. The comprehensive open-source will further reduce the development threshold of medical AI applications, allowing more medical institutions and developers to quickly develop medical AI applications tailored to their own needs. As one of the earliest companies in the Internet medical field to layout AI, JD HEALTH has been committed to long-term investment in AI capabilities and now has industry-leading self-developed medical large-scale models and representative medical AI products, achieving the largest scale of full-scene applications and being at the forefront of AI applications in the medical field. Since starting the Internet hospital business in 2018, JD HEALTH has closely integrated AI technology with business development, first launching innovative applications such as AI triage and AI prescription review. In July 2023, JD HEALTH released the self-developed medical large-scale model "Jing Yi Qian Xun," providing a solid technical foundation for the construction and upgrading of the medical service ecology. In early 2024, the JD HEALTH Exploration Research Institute (JDH XLab) was formally established to focus on cutting-edge AI research. Meanwhile, a series of AI products such as the AI assistant for doctors, "Chat Healing Universe," "Kangkang," and "Mobile Blood Pressure Measurement, have been successively launched for doctors and users. In January 2025, JD HEALTH officially released the full series of large-scale products based on the online universal platform, "AI Jing Yi," as well as the industry's first full-scene large-scale model product for hospitals, "JD Zhongyi." While building capabilities and developing products, how to promote the actual value of medical large-scale models in specific scenarios has always been a issue that JD HEALTH has been focusing on and working to solve. Take "AI Jing Yi" as an example. In this series of products, it includes the AI digital avatars of the first batch of domestic expert doctors, as well as a series of professional service DIGIHUMAN products such as AI pharmacists, AI nutritionists, and AI psychological consultants. These DIGIHUMAN not only have highly intelligent and personalized service capabilities, but also provide comprehensive health management and medical services according to the user's needs. As of now, the highest artificial replacement rate of JD HEALTH's medical AI DIGIHUMAN has exceeded 90%, with a service satisfaction rate of 91%. It has served over 35 million users in the past year, allowing users to enjoy the ultimate experience and professional security brought by AI 24/7. In addition, over 80% of the doctor's consultation forms in JD Internet Hospital use AI services, and the AI auxiliary diagnosis products like "Intelligent Doctor's Assistant" have become the smartest diagnostic tools with the highest usage rate and adoption rate among doctors. This has effectively helped doctors reduce the risk of misdiagnosis and free up more clinical energy, allowing doctors to focus on more valuable work tasks. Medical AI has both enormous social value and commercial potential. Wang Guoxin, Chief Scientist of JD HEALTH Exploration Research Institute (JDH XLab), stated that medical AI is not only an efficiency tool but also a new productive force reshaping the medical industry, and will bring significant value in solving the supply of medical resources. From technological breakthroughs, to large-scale implementation, and to ecological co-construction, JD HEALTH will take the open-source launch of the "Jing Yi Qian Xun" medical large-scale model as a new starting point, working with industry partners to accelerate the development of the medical AI industry ecosystem and promote the accessibility of high-quality medical resources.