研究背景 近年来，人工智能技术迅猛发展，大模型作为新一代通用智能技术底座，正深刻重塑医疗健康行业的服务模式与技术路径。 医疗大模型 涵盖医学影像、诊疗辅 助、医院管理、药物研发、健康管理、中医药等多个细分领域，展现出强大的技术赋能潜力与广阔的市场前景。 党中央、国务院高度重视人工智能与医疗健康的融合发展，《 "十四五"数字经济发展规划》《"人工智能+"行动方案》等政策文件明确提出推动人工智能在医疗领域 的深度应用。医疗大模型作为"技术+医疗"融合的典型代表，不仅能提升诊疗效率、优化医疗资源配置，也为构建普惠、精准、智能的新型医疗服务体系提供了关键 支撑。在此背景下，系统梳理医疗大模型产业生态、研判技术演进趋势、挖掘典型应用场景与商业模式，对推动行业从"技术探索"走向"规模化落地"具有重要战略意 义。 征集 目的 艾瑞研究院计划于 202 6 年发布《中国医疗大模型行业研究报告》，全面梳理行业发展脉络，洞察技术演进与市场趋势，展示行业优质厂商的创新实践与落地成 果。报告旨在为医疗机构 、技术企业、投资机构及政策制定者提供决策参考，推动行业健康、有序、高质量发展。 在此背景下，艾瑞咨询研究院正式启动医疗大模型产业 ...

健康管理平台和解决方案（HMPS）新增河北和广州两地"惠民保"业务。 机构：光大证券 研究员：付天姿/王贇 事件：近期公司在AI 医疗创新领域取得多项进展，包括深度参与北京市医疗领域国家人工智能应用中 试基地、中标III 期临床研究项目、新增河北和广州两地"惠民保"业务。 深度参与中试基地建设，以核心技术共筑AI 医疗创新高地。9 月12 日，公司作为北京市医疗领域国家 人工智能应用中试基地的核心生态伙伴，深度参与平台建设与场景创新，推动多项标杆项目落地：开发 区域级受试者招募AI智能体，全面提升受试者招募效率与精准度，解决临床研究受试者招募难、招募 慢的核心痛点，已在6 个项目中落地验证；开发北京市AI 应用插件程序，以统一入口方式整合集成多 AI 应用，大幅减轻医疗机构用户操作和系统集成负担。2025 年8 月，国务院发布《关于深入实施"人工 智能+"行动的意见》，文件指出，要优化应用发展环境，布局建设一批国家人工智能应用中试基地。 而后，上海、北京、广州、浙江相继宣布启动医疗领域国家人工智能应用中试基地，我们认为公司基于 医疗大模型技术将深度受益。 生命科学解决方案（LSS）中标III 期临床研究项目， ...

2025 年 11 月 23 日 公司研究 AI 医疗创新领域取得多项进展，医疗大模型构建"数据-算法-场景"飞轮闭环 ——医渡科技（2158.HK）跟踪点评报告 要点 事件：近期公司在 AI 医疗创新领域取得多项进展，包括深度参与北京市医疗 领域国家人工智能应用中试基地、中标 III 期临床研究项目、新增河北和广州 两地"惠民保"业务。 深度参与中试基地建设，以核心技术共筑 AI 医疗创新高地。9 月 12 日， 公司作为北京市医疗领域国家人工智能应用中试基地的核心生态伙伴，深度参 与平台建设与场景创新，推动多项标杆项目落地：开发区域级受试者招募 AI 智能体，全面提升受试者招募效率与精准度，解决临床研究受试者招募难、招 募慢的核心痛点，已在 6 个项目中落地验证；开发北京市 AI 应用插件程序， 以统一入口方式整合集成多 AI 应用，大幅减轻医疗机构用户操作和系统集成 负担。2025 年 8 月，国务院发布《关于深入实施"人工智能+"行动的意见》， 文件指出，要优化应用发展环境，布局建设一批国家人工智能应用中试基地。 而后，上海、北京、广州、浙江相继宣布启动医疗领域国家人工智能应用中试 基地，我们认为公司 ...

Core Insights - The rapid development of artificial intelligence technology, particularly large models, is significantly transforming the healthcare industry, enhancing service models and technological pathways [2] - The Chinese government emphasizes the integration of AI and healthcare, with policies promoting deep applications of AI in the medical field, aiming to improve diagnostic efficiency and optimize resource allocation [2] - The systematic review of the medical large model industry ecosystem and the exploration of typical application scenarios and business models are crucial for transitioning the industry from "technical exploration" to "scaled implementation" [2] Research Objectives - The iResearch Institute plans to release the "China Medical Large Model Industry Research Report" in 2026, aiming to provide insights into industry development, technological evolution, and market trends [3] - The report will serve as a decision-making reference for medical institutions, technology companies, investment institutions, and policymakers, promoting healthy and orderly industry development [3] Research Content - Define the concept, classification, and development history of medical large models, highlighting their strategic value in the intelligent transformation of healthcare [5] - Analyze the industry chain structure of medical large models and examine their application modes in areas such as medical imaging, clinical assistance, drug development, health management, and traditional Chinese medicine [5] - Assess the commercialization pathways and future development directions of medical large models, providing forward-looking strategic references for industry stakeholders [5] Participation Value - Companies participating in the report will have the opportunity to be included in the iResearch "2026 China Medical Large Model Industry Research Report" industry map and case studies, enhancing brand influence and industry recognition [7] - The report will be published on iResearch's official website and WeChat account, with dissemination through various media channels linked to iResearch [7] - Participants will be invited to attend iResearch's online and offline events in the healthcare and digital intelligence sectors, facilitating in-depth exchanges with industry experts and investment institutions [8] Target Companies - Both listed and unlisted companies engaged in the research, application, or service of medical large models are eligible, including those in medical imaging, clinical assistance, hospital management, drug development, health management, and traditional Chinese medicine [10] - Companies should have at least two years of experience in the relevant field and possess multiple practical case studies demonstrating industry representation and exemplarity [10]

Core Insights - Hangzhou Quanzhen Medical Technology Co., Ltd. has completed a Series B financing round of 100 million RMB, with strategic investments from Innovation Medical and Good Doctor Group expected in 2024 and 2025 respectively [2] - The company's flagship product, Trizen AI, has seen a tenfold increase in contract amounts with hospitals from January to October 2025 compared to the entire year of 2024, with an expected year-on-year growth of 12-15 times [2] Company Overview - Quanzhen Medical is one of the earliest companies in China focusing on post-training and clinical application of medical large models, aiming to enhance doctors' capabilities through AI [2] - The founder, Dr. Xue Chong, has a dual background as a clinical doctor and AI researcher, emphasizing that AI should assist rather than replace doctors [2] Product and Technology - Trizen AI is the first intelligent medical record and clinical assistance system in China, featuring automatic medical record generation, clinical decision support, and patient service [5] - The system improves medical record writing efficiency by over 65% and supports various types of medical documents, outperforming similar products in accuracy and language logic [5] - Quanzhen Medical is expanding its product from a single medical record tool to a comprehensive "AI doctor assistant" platform, covering the entire patient care process [5] Technical System - The company has established a complete "foundation-model-application" technical system since 2023, focusing on post-training methods tailored for medical scenarios [6] - Trizen AI utilizes localized large model deployment, compressing general models into specialized medical models with 1 to 10 billion parameters, significantly reducing inference costs and improving response speed [9] Commercialization - By the third quarter of 2025, Trizen AI has been widely implemented, reducing inpatient medical record writing time from an average of 10 minutes to 1.8 minutes in some hospitals, with a diagnostic accuracy rate of 93.5% [12] - The company has developed specialized models for different medical fields, including traditional Chinese medicine and pediatrics, enhancing the digital transformation of hospitals [13] Conclusion - Quanzhen Medical represents a new phase in China's medical AI landscape, transitioning from algorithm experimentation to clinical application, aiming to integrate AI capabilities with medical practice [14][15]

Core Insights - The core viewpoint of the article highlights the launch of iFlytek's medical large model, specifically the Smart Medical Assistant Hospital Version 1.0, which demonstrates significant advancements in diagnostic accuracy and efficiency in complex inpatient scenarios [1] Company Summary - iFlytek has released the Smart Medical Assistant Hospital Version 1.0, which is designed for tiered hospitals and aims to enhance medical diagnosis and treatment [1] - The upgraded version of iFlytek's Xiaoyi has also been introduced alongside the new model [1] Performance Metrics - The specialized AI capabilities of the iFlytek medical large model have reached the level of chief physicians in tiered hospitals for the first time in the industry [1] - Pilot data indicates that the specialized diagnostic rationality rate has improved to 96%, while the cross-departmental diagnostic rationality rate has reached 91% [1] - The time required for medical record writing has been reduced by 50% [1]

Core Insights - Ant Group's CEO announced a significant organizational restructuring, elevating the "Digital Healthcare Division" to a "Healthcare Group," aiming to accelerate the healthcare business as a strategic pillar of the company [6][10] - This restructuring marks a critical move for Ant Group as it seeks new growth avenues, completing its AI strategic layout across three key sectors: lifestyle services, financial services, and healthcare services [7][9] Organizational Changes - The newly formed Healthcare Group will join four other core business units: Alipay Group, Digital Payment Group, Wealth Insurance Group, and Credit Group [6] - Zhang Junjie, a veteran with Ant Group since 2014, has been appointed as the president of the Healthcare Group, bringing extensive experience in the company's healthcare initiatives [10][11] Market Potential - The Chinese healthcare market is projected to exceed 20 trillion yuan by 2025, indicating a vast opportunity for growth [9] - The aging population in China is driving significant demand for chronic disease management and health care services, making this an opportune time for Ant Group to invest heavily in this sector [9] AI Integration - Ant Group's AI Health Manager AQ has rapidly gained traction, becoming the first industry-specific AI application in China to surpass 10 million monthly active users [8][11] - The company's medical AI model has received high ratings in various domestic and international healthcare assessments, showcasing its capability to address core medical scenarios [8][11] Strategic Focus - The elevation of the healthcare sector to a strategic level reflects Ant Group's proactive approach to addressing societal healthcare challenges through technology [8][12] - The company emphasizes its commitment to leveraging AI as a key pathway to solve pressing healthcare issues, aligning with its mission to use technology to address social problems [12][13]

Core Insights - The latest technology upgrade of iFlytek's Spark model, named Spark X1.5, was announced at the 8th World Voice Conference, showcasing advancements in deep reasoning capabilities and multi-language support exceeding 130 languages [1][3] - iFlytek's medical AI model has achieved significant breakthroughs, particularly with the launch of the Smart Medical Assistant Hospital Edition 1.0, enhancing diagnostic capabilities and efficiency in healthcare settings [3][6] Group 1: Technology Advancements - iFlytek has successfully implemented efficient long-term training of the MoE model on a fully domestic computing power platform, achieving performance comparable to leading international models with half the parameters [1] - The Spark medical model has surpassed GPT-5(high) and DeepSeek V3.2-Exp in various capabilities, including medical knowledge Q&A, language understanding, and multi-modal interactions [3][4] Group 2: Medical AI Applications - The Smart Medical Assistant can now support diagnosis for 2,500 diseases with a first diagnosis accuracy of 98%, integrating evidence-based reasoning and traditional Chinese medicine capabilities [3][4] - In outpatient scenarios, the diagnostic accuracy of the Spark medical model reached 93.1%, outperforming both junior and senior physicians, while in inpatient settings, it achieved a specialist-level performance [4][5] Group 3: User Engagement and Impact - The Smart Medical Assistant has provided 1.1 billion auxiliary diagnoses across 801 districts, addressing the dual challenges of efficiency and quality in healthcare [6][12] - The upgraded iFlytek Xiaoyi app has completed over 160 million AI health consultations with a user satisfaction rate of 98%, offering personalized health management services [9][10] Group 4: Industry Context and Future Directions - The Chinese government has issued guidelines to promote the integration of AI in healthcare, positioning iFlytek as a key player in this initiative through continuous technological innovation and application [12] - iFlytek aims to expand the boundaries of medical AI innovation and application depth, aligning with national strategies for high-quality development in the healthcare sector [12]

Core Insights - Baichuan Intelligent has announced the release of the evidence-enhanced medical model Baichuan-M2Plus, along with an upgrade to its application Baixiao Ying and the opening of its API [1] - This release follows the open-sourcing of Baichuan-M2 in August, marking another significant step for the company [1] - Evaluation results indicate that M2Plus has a significantly lower medical hallucination rate compared to general models, approximately three times lower than DeepSeek, and it performs better than the U.S. medical product OpenEvidence, achieving a credibility level comparable to that of experienced clinical doctors [1]

Policy Developments - The Central Commission for Discipline Inspection and the National Health Commission have issued an implementation plan to strengthen supervision over major national health science and technology projects, focusing on key stages such as project formation, evaluation, management, performance assessment, and fund usage [2] Drug and Device Approvals - Jiukang Bio has received a medical device registration certificate for its digoxin testing kit, which enhances its product line in the in vitro diagnostic field, although future sales depend on market promotion effectiveness [4] Financial Reports - Hongyuan Pharmaceutical reported a third-quarter revenue of 435 million yuan, a year-on-year increase of 2.45%, and a net profit of 4.71 million yuan, reflecting a significant year-on-year growth of 1,316.05% [6] - Dawei Medical disclosed a third-quarter revenue of 141 million yuan, a year-on-year increase of 3.13%, with a net profit of 21.64 million yuan, up 28.96%. For the first three quarters, revenue was 393 million yuan, down 2.55%, while net profit increased by 14.71% to 72.62 million yuan [7] Capital Market - Jiankang Medical has indirectly held approximately 0.05% of shares in Zhiyuan Innovation (Shanghai) Technology Co., Ltd. through its investment in a fund [9] Industry Events - Jitai Technology announced that its AI-driven drug MTS-004 has completed Phase III clinical trials, becoming the first AI-enabled formulation drug in China to reach this milestone, with plans to submit a new drug application in 2026 [13] - Baichuan released the evidence-enhanced medical large model Baichuan-M2 Plus, which significantly reduces the rate of medical hallucinations compared to general models [14] - A new treatment for Helicobacter pylori infection has been developed by a team from Peking University Third Hospital, offering a simpler triple therapy option [15] Public Sentiment Alerts - Qianyuan Pharmaceutical announced that major shareholder Weng Zanguo plans to reduce his stake by up to 3% within three months due to personal financial needs [17]