iFlytek Spark Medical LLM V3.7 upgraded; AI Patient Management 3.0 released and industry standard co-development launched

On October 22, the main forum of Iflytek Co., Ltd.'s Global 1024 Developer Festival was held at the Beijing National Convention Center. At the event, iFlytek Spark Medical LLM V3.7 was officially released, achieving multimodal technology breakthroughs in deep integration of medical semantics, imaging, and speech based on a fully domestic computing power foundation, marking that iFlytek Healthcare's (02506) AI capabilities have officially crossed the threshold of clinical practicality and are being comprehensively and deeply integrated into the entire medical chain. iFlytek Xiaoyi was simultaneously upgraded with capabilities including comprehensive interpretation of multiple medical reports, health risk insights and comprehensive health management plans, and TCM tongue and facial constitution identification and wellness plans, providing individuals and industry clients with higher-quality full-cycle health management services. At the AI Patient Management Innovation Forum held concurrently, iFlytek Healthcare released the AI Patient Management Platform 3.0 and, together with domestic tertiary Grade-A hospitals and authoritative research institutions, launched the AI Patient Management Expert Consensus Special Program, leading the standardized and intelligent upgrade of the AI patient management industry from three major dimensions: technological breakthroughs, product iteration, and standard co-development. Spark Medical LLM V3.7 Upgraded, Multimodal Empowerment for All Medical Scenarios As the core technical achievement in the medical field at this Developer Festival, Spark Medical LLM V3.7 achieved all-round capability advancement, leading comprehensively in core capabilities such as medical knowledge Q&A, clinical language understanding, treatment plan recommendations, medical document generation, multimodal interaction, and multi-turn evidence-based reasoning. In the latest MedBench5.0 agent evaluation released by Shanghai Artificial Intelligence Laboratory, the model topped the industry ranking with a comprehensive score of 99.6 points, with its core technical strength receiving authoritative certification. Relying on the newly upgraded multimodal technical capabilities, iFlytek Healthcare simultaneously achieved deep empowerment in both doctors' clinical diagnosis and treatment and residents' full-cycle health management scenarios. In the field of AI diagnosis and treatment assistants, the model's end-to-end intelligent medical record generation reasonableness rate increased to 85%, completing the experience leap from "usable" to "easy to use"; medical imaging-assisted diagnosis capabilities improved by more than 15% compared with industry vertical models, and automated report generation and deep quality control capabilities reached industry-leading levels, effectively helping hospitals improve the quality and efficiency of imaging diagnosis and treatment. In the field of AI health assistants, iFlytek Xiaoyi completed all-round upgrades including intelligent interpretation of medical reports, dynamic health record management, personalized risk warnings, TCM constitution identification, and customized health plan generation, driving the service to iterate from one-time consultation to continuous, personalized full-cycle health management. The continuously iterating technology foundation supports large-scale implementation of products in multiple scenarios, accumulating massive real-world medical service data and mature evidence-based capabilities. Currently, the Smart Medical Assistant has completed more than 1.2 billion assisted diagnoses and 490 million assisted medical record writings; the imaging cloud platform provides an average of 60 million services per year and stores 160 million images; iFlytek Xiaoyi's total AI health consultation services exceed 220 million, with downloads exceeding 44 million and user satisfaction reaching 98%, laying a solid technical and data foundation for deepening patient management. Tertiary Grade-A Benchmark Leaders Pioneer Practical Exploration, Letting Innovation Empower Medical Staff and Serve Patients On top of the continuously iterating technology foundation and large-scale implementation capabilities, iFlytek Healthcare has joined hands with multiple tertiary Grade-A hospitals to take the lead in carrying out exploratory patient management practices. This AI Patient Management Innovation Forum, adhering to the industry original aspiration of solving post-diagnosis management pain points, making up for shortcomings in continuity of medical services, and benefiting more patients, gathered forces from industry, academia, research, and application to jointly explore development paths. Leaders and experts from the National Health Commission Medical Administration Center, China Academy of Information and Communications Technology, China Cardiovascular Health Alliance, China-Japan Friendship Hospital, the National Clinical Research Center for Orthopedics and Sports Rehabilitation of the Chinese PLA General Hospital, Shandong University Qilu Hospital, the First Affiliated Hospital of University of Science and Technology of China (Anhui Provincial Hospital), Xiangya Hospital of Central South University, Xuanwu Hospital of Capital Medical University, Peking University First Hospital, Beijing Anding Hospital, as well as the Chinese Medical Association Publishing House and People's Medical Electronic Audio-Visual Publishing House, attended. The forum was moderated by Ge Lei, Vice President of iFlytek Healthcare and General Manager of the Patient Management Product Line. At the forum, representatives from three hospitals respectively presented distinctive implementation practice results, using real clinical data to introduce the practical value of AI patient management and accumulating replicable and promotable practical experience for the entire industry. Tian Yanghua, President of the First Affiliated Hospital of University of Science and Technology of China (Anhui Provincial Hospital), stated that the hospital jointly built the "Smart Doctor Companion" large model with iFlytek Healthcare, connecting the full-chain scenarios of intelligent pre-diagnosis guidance, full-process accompaniment during diagnosis, and post-diagnosis patient management, adopting a joint model of "medical staff focusing on core diagnosis and treatment + platform professional operational support," providing standardized basic services for all hospital patients and refined management for key specialty disease populations. It has now been launched in more than 30 clinical departments, covering more than 100 diseases, with online consultation services exceeding 180,000 person-times and patient satisfaction reaching 97%. Specialty projects such as cerebrovascular intervention, integrated maternal and perinatal care, and weight management have achieved remarkable results. Han Hui, Vice President of Shandong University Qilu Hospital, stated that the hospital jointly released the "Qilu Spark" full-course management large model with iFlytek Healthcare. Relying on the human-machine collaborative model of "AI preliminary assessment + health manager intervention + doctor review," and through tiered and classified management, it has covered 20 clinical departments and 70 diseases, cumulatively managing more than 270,000 patient person-times free of charge, carrying out specialty disease follow-ups for 27,000 person-times, answering more than 100,000 consultation messages, and reducing medical staff's follow-up workload by 90%; patient satisfaction exceeds 98%, the planned return visit rate reaches 70%, and medication compliance reaches 97%. At the same time, Yuan Ye, an expert from Xiangya Hospital of Central South University, also shared the hospital's excellent practices in full-course case management. AI Patient Management 3.0 Freshly Released, Gathering Forces to Build New Industry Standards On the basis of concentrated sharing of benchmark practices and building industry consensus, this forum further promoted implementation of results and industry co-development: iFlytek Healthcare productized and standardized the pioneering exploration experience of multiple leading hospitals and officially launched the AI Patient Management Platform 3.0; at the same time, it joined with all industry parties to launch the co-development of expert consensus and standard systems, allowing pioneering experience to benefit a broader range of medical institutions and patient groups. The AI Patient Management Platform 3.0 completed systematic upgrades with "new scenarios, new AI, and new experience." At the scenario level, the product expanded from single-department specialty disease management to hospital-wide tiered services, having accumulated more than 160 standardized specialty disease management pathways and covering more than 30 mainstream clinical departments; at the AI capability level, relying on more precise patient profiles, intelligent follow-up agents, and dynamic evidence-based reasoning capabilities, it can automatically adjust management plans according to patients' rehabilitation progress, intelligently identify risks, and trigger tiered interventions. Combined with the multi-specialist co-management model of doctors, nurses, nutritionists, and rehabilitation therapists, it realizes the upgrade from fixed mechanical follow-ups to individualized, continuous, and precise full-course care; at the user experience level, it achieves centralized service entrances, clear tasks, and concrete guidance on the patient side, and aggregated work to-dos, efficient collaboration, and traceable processes on the medical staff side, greatly reducing the burden of out-of-hospital management work. In response to the current industry situation in AI patient management of no unified norms, no standard pathways, and uneven service quality, this forum officially launched the AI Patient Management Expert Consensus Special Program. iFlytek Healthcare will join hands with national benchmark hospitals, authoritative academic institutions, and industry experts to systematically sort out the service processes, technical specifications, quality control standards, and application boundaries of AI patient management based on massive real-world clinical service cases, fill the gap in industry standards, and build a scientific, standardized, and implementable new paradigm for AI patient management. Shandong University Qilu Hospital, Chinese Medical Association Publishing House, the First Affiliated Hospital of USTC (Anhui Provincial Hospital), China Cardiovascular Health Alliance, iFlytek Healthcare, and others jointly took the stage to participate in the launch Tao Xiaodong, President of iFlytek Healthcare, stated that the construction of Healthy China requires continuous strengthening of national health management, disease prevention and control, and continuous medical service capabilities. Relying on independently controllable domestic AI computing power and medical large models, iFlytek Healthcare continues to extend its capabilities from in-hospital digital and intelligent diagnosis and treatment to out-of-hospital patient management, insisting on starting from real diagnosis and treatment needs, empowering medical staff with technology, and serving patients with products. The core goal is to make up for shortcomings in the continuity of medical services and promote the true transformation of medical services from "disease-centered" to "patient-centered." Currently, the AI Patient Management Platform has been implemented in more than 300 hospitals nationwide, cumulatively completing refined management of chronic diseases and high-risk populations for more than 500,000 person-times. Relying on the dual drive of an authoritative medical knowledge base and real-world big data, the platform builds a solid evidence-based foundation with an authoritative medical knowledge base on the scale of nearly 1 billion, and combines with real service data of more than 2.2 million person-times per day on both the doctor and patient sides to continuously feed model iteration, forming a positive cycle of technology iterationproduct implementationclinical validationstandard improvement. Mature in-hospital specialty disease management pathways are entering thousands of households through iFlytek Xiaoyi. These pathways, verified by large-scale clinical results, are transformed into executable and trackable daily health plans on everyone's mobile phone, connecting medication reminders, re-examination follow-ups, and diet and exercise guidance, and dynamically adjusting with changes in indicators. The deeper the hospital cooperation, the more professional the personal health management services becomehigh-quality medical capabilities thus break through hospital walls and extend into residents' daily lives, allowing more people to receive intervention and management before disease occurs. From leading the industry in large model technology, to large-scale implementation of all-scenario products, to taking the lead in co-building industry standards, iFlytek Healthcare has always used AI technology to solve the industry pain points of patient management, making a pragmatic response to strengthening national health management capabilities under the Healthy China strategy. In the future, iFlytek Healthcare will continue to use the Spark Medical LLM as its technology foundation, deepen the construction of a full-chain smart medical service system covering in-hospital diagnosis and treatment, out-of-hospital follow-up, grassroots linkage, and individual proactive management, and through technological innovation, product upgrades, and standard co-development, promote the upgrade of medical and health services from standardized universal access to personalized precision, help high-quality medical resources sink to lower levels, and enable continuous, warm full-course health services to benefit broader populations.