Recently, Baichuan's large model officially launched its latest medical large model - Baichuan-M2Plus, and upgraded the accompanying application BaiXiaoYing at the same time, opening up API interfaces. This release marks another important progress for Baichuan after launching the open-source model Baichuan-M2.
Through a series of evaluations, M2Plus has shown excellent performance in reducing medical hallucination rates, significantly lower than general large models. Especially when compared with existing medical products such as DeepSeek, the hallucination rate was reduced by about three times, even surpassing the popular medical application OpenEvidence in the United States.
Image source note: The image was generated by AI, and the image licensing service provider is Midjourney
M2Plus adopts the six-source evidence-based reasoning (EAR) paradigm, becoming an intelligent assistant known as the "Doctor's ChatGPT." This model addresses application challenges in serious medical scenarios, integrating original research, evidence reviews, guidelines, practical knowledge, public health education, and real-world information from regulations to build a complete medical knowledge system, ensuring the credibility and scientific nature of medical decisions.
In terms of evidence-based retrieval, M2Plus uses the PICO framework to transform medical queries into structured questions, ensuring that the retrieved information is accurate and reliable. The design of this model allows doctors to obtain high-level, credible medical evidence when facing complex medical issues, greatly improving the efficiency of using medical information.
More notably, when answering medical questions, M2Plus employs an "evidence-enhanced training" mechanism, ensuring that the model's answers are not only based on retrieved evidence but also effectively avoid generating arbitrary information. By reinforcing references to authoritative sources and evaluating evidence quality, M2Plus demonstrates credibility comparable to that of experienced clinical experts.
In multiple medical scenario tests, M2Plus has received high praise from clinical doctors, especially in analyzing medical history, diagnostic thinking, and treatment plans, where the accuracy and professionalism of its answers have been widely recognized. Additionally, in the United States Medical Licensing Examination (USMLE), M2Plus achieved an impressive score of 97, further proving its application potential in the medical field.
Key Points:
🌟 Launch of the M2Plus model, significantly reducing medical hallucination rates and surpassing multiple existing medical products.
🔍 Adopting the six-source evidence-based reasoning (EAR) paradigm to ensure the scientific and credible nature of medical decisions.
🏆 Achieved a score of 97 in the USMLE exam, demonstrating outstanding medical professional capabilities.
