Ladies, Your AI Chief Physician Has Arrived! China's First Gynecology-Specific Large Model Passes Senior Professional Title Exam

On July 28, Doukou, a gynecological large model developed by Yisheng Jiankang (Hangzhou) Life Science & Technology Co., Ltd., successfully passed the national senior professional title (Level I Senior) written examination in obstetrics and gynecology. This achievement marks it as the first vertical medical AI model trained on DingTalk’s enterprise-exclusive AI platform by a startup company to meet such a high-level professional standard. This breakthrough not only signifies a key leap forward in vertical large-model development within the healthcare sector but also provides the industry with a replicable innovation pathway through its exemplary impact.

Technological Breakthrough and Efficient Development Pathway

From technical攻坚 to practical implementation, the Doukou gynecological large model achieved rapid development—from concept to optimized performance—within a short timeframe, meeting stringent professional qualification standards. This efficient progress challenges the conventional belief that developing vertical large models requires massive investment and long development cycles. It demonstrates that small-to-medium-sized teams can leverage professional training platforms, scientific methodologies, high-quality domain-specific data, and focused technical strategies to quickly build proprietary large models reaching top-tier professional levels.

Core Technical Support and Training Methodology

The Doukou model's success was made possible by starting with an advanced foundational model and leveraging DingTalk’s enterprise-exclusive AI platform and specialized services. By constructing high-quality obstetrics and gynecology datasets and applying multi-stage optimization techniques, the team significantly enhanced model performance.

Zhu Hong, CTO of DingTalk, stated that Doukou is the first professional vertical large model born on DingTalk’s AI platform. Through just over one month of collaboration, the joint team boosted the model’s accuracy to 90.2%, enabling it to pass the professional exam. “This validates DingTalk’s capability to help enterprises across industries build their own proprietary large models. We are continuously improving our support system for industry- or enterprise-specific large models and pioneering an AI model pay-per-performance model, empowering more industry-focused companies like Yisheng Jiankang to truly implement AI applications,” Zhu said.

Throughout the development process, Yisheng Jiankang and DingTalk adopted a technical approach combining “high-quality, precisely annotated medical data + customized training tools + efficient training workflows and methods.” This enabled rapid iteration of the Doukou model while significantly improving its accuracy and stability, allowing strong performance even in complex clinical scenarios. The model’s journey—from data preparation and preprocessing to continuous performance optimization—offers a replicable reference case for building specialized large models in healthcare and beyond.

Examination Standards and Evaluation Results

The National Senior Professional Title Examination (Level I Senior) in Obstetrics and Gynecology is the gold standard for assessing the expertise of OB-GYN physicians. Covering 12 core disciplines—including female reproductive anatomy, clinical obstetrics and gynecology, and reproductive endocrinology—the exam rigorously evaluates real-world capabilities such as diagnosing complex cases and designing high-difficulty surgical plans, demanding the kind of "clinical intuition" typically gained from decades of practice.

This evaluation strictly used the People's Medical Publishing House edition of the *Full-Scale Simulation Test for the Senior Professional Title Examination in Obstetrics and Gynecology*, designated by the National Health Commission. The test covered 12 core areas including clinical obstetrics and gynecology, gynecologic oncology, perinatal medicine, reproductive endocrinology, and family planning. The format included multiple-choice questions (40%) and case analysis questions (60%). The case analysis section required the model to integrate patient complaints, lab reports, and other multi-source information to address clinical diagnosis, differential diagnosis, and treatment planning—comprehensively testing clinical decision-making skills. Full correctness was required to score points, making the evaluation criteria stricter than those applied in actual human exams.

Multiple-choice accuracy: 75.56%; Case analysis (multiple-select) accuracy: 59.01%; Final overall accuracy: 64.94%. The model outperformed several other models on both question types. To ensure reliability, the team validated results using three independent test papers and calculated the average score.

(Comparison based on the same exam paper)

Application Prospects and Industry Impact

"Passing the Level I Senior exam means the model has attained the professional judgment level of a chief physician," emphasized Wang Qiangyu, founder of Yisheng Jiankang. "Our practice proves that even small and medium enterprises can train highly accurate large models." However, Wang also stressed that large models will not replace OB-GYN doctors. Their core value lies in: providing home-based self-diagnosis support for women, enabling "pre-visit triage" and "out-of-hospital health management"; offering science communication guidance and lifestyle recommendations for non-urgent cases; and delivering expert-level support to medical and aesthetic institutions to improve gynecological service quality. Additionally, by training specialty models on institutional data, patients can receive pre-consultation services at the level of top specialists, enhancing healthcare efficiency.

Industry experts noted: "This breakthrough opens new pathways for deep AI applications in clinical decision support, evidence-based medicine research, patient education, and medical student training in obstetrics and gynecology." Dr. Zhou from the Women's Hospital, School of Medicine, Zhejiang University, gave high praise: "This advancement will greatly facilitate our work and help improve diagnostic efficiency and accuracy."

As the technology continues to evolve and expand, the Doukou gynecological large model is expected to play an increasingly important role across diverse medical settings. It will further optimize the allocation of healthcare resources and help alleviate the uneven distribution of high-quality gynecological care. In the future, the model will collaborate with more medical institutions to drive intelligent and efficient development in the healthcare industry, bringing tangible benefits to more women patients.

We dedicated to serving clients with professional DingTalk solutions. If you'd like to learn more about DingTalk platform applications, feel free to contact our online customer service or email at This email address is being protected from spambots. You need JavaScript enabled to view it.. With a skilled development and operations team and extensive market experience, we’re ready to deliver expert DingTalk services and solutions tailored to your needs!