JD Health's 'Jingyi Qianxun' AI Model Tops MedBench Ranking, Accelerating Large Model Applications in Healthcare

On July 15, MedBench, an open evaluation platform for Chinese large language models in the medical field, updated its leaderboard. JD Health’s “Jingyi Qianxun” large model topped the list with an outstanding overall score of 92.4, demonstrating its leading capabilities in the realm of intelligent healthcare.

Assessment Chart

As an authoritative evaluation platform launched by the Shanghai AI Laboratory and the Shanghai Digital Medical Innovation Center, MedBench leverages the expert experience and knowledge reserves of top-tier medical institutions to build a fair and rigorous evaluation system for Chinese large language models in healthcare. Currently, MedBench encompasses five major evaluation dimensions: medical language understanding, medical language generation, medical knowledge question answering, complex medical reasoning, and healthcare safety and ethics. It includes 300,000 specialized Chinese medical evaluation questions, providing objective and scientific performance benchmarks for Chinese large language models in the medical field.

In the evaluation, JD Health’s “Jingyi Qianxun” large language model demonstrated exceptional capabilities in handling complex medical scenarios. Particularly in core tasks such as medical knowledge question answering, case analysis, and medical record generation, the model exhibited superior accuracy and practicality compared to conventional models. In specialized assessments targeting specific medical tasks, including medical knowledge question answering and medical language understanding, the “Jingyi Qianxun” medical large language model ranked first, underscoring its profound expertise in the medical field.

It is reported that the “Jingyi Qianxun” medical large language model, released in July 2023, is the industry’s first healthcare LLM integrating knowledge with data. It has accumulated over 100 million high-quality health records covering both online and offline doctor-patient scenarios, and possesses extensive big data on the end-to-end circulation of pharmaceutical products.

Built upon the JD Yanxi general-purpose large language model, the “Jingyi Qianxun” large model integrates general data, healthcare-specific data, and native data from intelligent supply chains. With strong industrial relevance and generalization capabilities, it can rapidly adapt to and learn from various scenarios in the healthcare sector, thereby enabling comprehensive AI integration across products and solutions. Leveraging this foundation, JD Health has positioned “Jingyi Qianxun” as the “technical backbone” for building a new generation of healthcare service ecosystems, driving the intelligent upgrade of industry-wide solutions and products.

Currently, the AI-assisted diagnostic accuracy rate of JD Health’s Dermatology Hospital, powered by large language models (LLMs), exceeds 95%. The AI-based pre-prescription review system, jointly developed by JD Health and the Haidian District of Beijing, has been deployed in nearly 100 primary healthcare institutions, processing over 20,000 prescriptions daily. Furthermore, leveraging the “Jingyi Qianxun” medical LLM, JD Health has launched a suite of solutions for licensed physicians, including cloud consultation rooms, post-consultation follow-ups, multi-specialist joint consultations, clinical research support, physician IP brand incubation, and the “Intelligent Physician Assistant” tool. These initiatives have optimized the efficiency and quality of online diagnosis and treatment while enhancing practice safety. These cases fully demonstrate the robust capabilities of the “Jingyi Qianxun” medical LLM in improving the quality and efficiency of healthcare services, enhancing patient care experiences, and driving digital transformation within the industry.

JD Health has long leveraged its technological expertise, supply chain capabilities, and medical service resources to drive high-quality development of the healthcare system through innovative, internet-based products, systems, and solutions centered on user experience. Its top ranking in the MedBench benchmark underscores JD Health’s significant advantages in foundational resources for industry-specific large language models. Moving forward, JD Health will continue to deepen its engagement in digital and intelligent healthcare, promoting diverse applications of large models across various healthcare scenarios to further enhance the quality of medical and health services.