A：The Spark image model technically realizes unified quality control of all images, and its quality exceeds the level of traditional expert quality control centers, and can realize cross-hospital image review within 3 seconds, which greatly improves the work efficiency of doctors. In addition, through intelligent recognition technology, it can accurately remind the lesions in the image, compared with other large model report generation and artificial intelligence applications in the market, the spark image large model of Xunfei Medical in the field of imaging has a clear lead.

A：It took more than four years from construction to operation of Anhui Province Image Cloud, while in Guangxi, it only took 20 days to complete the first hospital online, and within 90 days to achieve the delivery and operation conditions of all hospitals above the second level. This experience of rapid construction and operation has summed up a complete set of methodology, which enables Guangxi image cloud to complete the construction and operation quickly and achieve good use effect.

A：The secondary hospitals in Guangxi have achieved 100% access and operation capability, with an image upload volume of 10 million and an annual inspection volume of about 30 million. Basically, the full amount of image data has been uploaded to the cloud and started operation. The doctor's access rate is close to 40%, and the patient sharing rate has also increased.

A：On May 20, 2026, the official website of the National Medical Insurance Bureau highly affirmed the construction speed and effectiveness of Guangxi Image Cloud, and promoted its construction and operation mode in the second half of the year. So far, the model has been extended to Ningxia and other places to promote the construction and operation of the national digital image cloud.

A：By the end of 2025, the digital imaging cloud platform has served 36 million patients in total, and the growth is rapid, and it is expected to reach more than 0.1 billion by the end of the year. With the construction and operation of image clouds in more districts, counties and provinces, the number will continue to accelerate. We will continue to promote services based on large image models, such as auxiliary examination, auxiliary diagnosis and pathological and examination medical imaging services, and expand to pan-image-related digital interconnection and mutual recognition services, and strive to build the world's largest regional medical imaging and pan-image data platform.

A：In addition to radiological imaging, it is also planned to expand the image cloud to pan-imaging, including all examination and inspection items such as ultrasound, endoscopy and ECG, so as to gather objective data of patients, form a perfect resident health file and provide a basis for digital plus one service.

A：Through the industry-leading medical model ability and general AI ability, combined with voice technology and authoritative experts to develop a specialized post-diagnosis management path, recognized by the head hospital. The service coverage rate of discharged patients was significantly improved, the satisfaction rate was as high as 95%, and the quality of rehabilitation was effectively improved, and the recurrence rate and complication rate were reduced. In the first half of this year, hospital coverage and patient reach achieved rapid operational growth. Revenue from patient management services in sales increased by 200 to 22.77 million yuan. Four new operational hospitals were added. The number of patients served increased from 20000 to 35000 and is expected to reach 55000 by the end of the year.

A：AR patient management is mainly deployed in hospitals. Patient portraits are generated by using hospitalization data, and personalized management paths are generated according to portraits and clinical guidelines, including follow-up reminders, follow-up plans, medication guidance, etc. Various data in the rehabilitation process are automatically executed and collected through artificial intelligence to form an analysis and intervention mechanism to improve the rehabilitation effect and safety.

A：In the future, it will serve the government, doctors and patients more deeply. Through AI digital base, national health information platform, AI consultation and other means, it will improve the service ability of primary doctors, strengthen the role of AI diagnosis and treatment assistants, expand the service scope of AI health assistants for individuals, provide full cycle health management, and ensure the continuous improvement of service quality and ability.

A：The medical big model maintains its industry-leading advantages in the fields of knowledge question and answer, language understanding, diagnosis and treatment recommendation, document generation and multi-round interaction. The ability of the AI health assistant agent for residents is also continuously improving, especially in the interpretation of inspection forms and physical examination reports. It has a clear lead in the interpretation of inspection forms and physical examination reports, and has comprehensively upgraded its medical multimodal ability, taking the lead in breaking through the practical threshold, it has reached the industry-leading level in medical speech recognition, case generation, image report analysis and generation.

A：We support the improvement of large model capabilities through the first-line real-world scenario of diagnosis and treatment data. Specifically, it includes providing diagnosis and treatment advice for primary medical institutions, recording doctors' adoption and professional corrections by superior doctors, and forming feedback to improve the professional knowledge of large models; at the same time, feedback from high-level professional doctors contributes to the precipitation of full-process diagnosis and treatment data and the continuous improvement of large model capabilities.

A：In terms of growth drivers, the growth of AI diagnosis and treatment assistant business is due to the accelerated promotion of primary medical institutions and the application of hospital-assisted case generation and Junior College artificial intelligence follow-up brought about by the technological progress of large models. In AI patient services, with the promotion of the policies of the National Health Insurance Administration and the acceleration of image cloud construction in various regions, digital image services will achieve rapid growth and become an important growth point for AI health assistants. In addition, the state's investment and funding in the field of AI health care is also a key factor in business growth. The growth momentum of the image cloud business comes from the national promotion under the policy guidance of the National Health Insurance Administration, and the successful replication of the Anhui and Guangxi models has enabled public hospitals to basically achieve full coverage, with a hospital-side sharing rate of more than 40%, effectively saving health insurance costs and preventing duplicate examinations. At present, the state has clearly requested the promotion of image cloud services nationwide. It is expected that it will occupy more than 60% of the national market share in the future. There will be hundreds of millions of inspection fees brought about by operating service fee income every year, forming a clear source of funds and sustainable growth space. At the same time, the image cloud will also be extended to more areas such as blood tests and urine tests, further improving patient satisfaction and hospital efficiency.

A：The growth drivers of AI diagnosis and treatment assistants mainly come from two aspects: first, the national promotion of the primary market is accelerated, especially the communication with the leaders of the Health Commission and the national primary medical layout have promoted the rapid development of business; Second, in secondary and tertiary hospitals, the technological progress of large models has lowered the use threshold in the fields of AI-assisted diagnosis and multidisciplinary consultation, thus providing hospitals with practical application scenarios. In the future, we will continue to deepen cooperation with the government and hospitals, use the advantages of credibility to expand the market, and carry out customized service promotion according to the characteristics of different regions.

A：In the medical scenario, the advantages of the vertical model of Xunfei Medical are mainly reflected in two aspects. First of all, the basic medical thinking chain is constructed based on clear and verified medical knowledge such as textbooks and clinical guidelines, which provides a credible and verifiable basis for generating content. Secondly, by combining the real world data and using the probability generation ability of the large model, the clinical medical thinking chain is constructed on the basis of the continuously upgraded base large model. This thinking chain provides personalized services under the scenes of doctor diagnosis and treatment and residents' health management consultation, and reduces the generation rate of illusion and improves the interpretability through the restriction of the basic medical thinking chain.

A：The key to building a high-quality data flywheel is to form a closed-loop feedback mechanism that contains one-line data. Early data comes from rigorous medical education materials and historical data from partner hospitals, while at present it is more important to combine millions of daily diagnostic and auxiliary diagnostic data with actual clinical feedback from doctors to form high-quality real data backflow. At the same time, the company is working with the top 100 hospitals in the country to collect high-level expert experience and in-depth data on post-diagnosis patient management to enhance the advantages of the data flywheel.

A：The advancement of the company's large model technology has promoted several levels of business development: one is to combine two thinking chains to explain the diagnosis and treatment process and provide more interpretable output; the other is to gradually cover services from the grassroots to the secondary and tertiary hospitals. Junior College, and even more subdivided areas of Junior College; the third is to provide more personalized services in residents' health consultation and health management through the combination of the two thinking chains, ensure that the generated health management content is highly compatible with residents' health demands.

A：The company has been committed to becoming "the first stock of China's medical model" and has maintained a leading position in many technical indicators. For example, it is the only robot that has passed the national professional physician qualification examination organized by the National Health Commission. In addition, the company outperformed other competitors in terms of diagnostic and treatment recommendation accuracy, reaching the level of Junior College doctors in top three hospitals.

A：The service package products have achieved significant value improvements in clinical applications, such as an increase in patient satisfaction to 98%, a reduction in unplanned readmission rates from 1.9 to 0.9, and a reduction in complication rates to zero. For doctors, the product helps doctors find high-risk patients in time and intervene in management through personalized path setting and automatic execution, thus improving the quality of patient rehabilitation and the efficiency of doctor management without adding too much workload. For the expansion plan in the next three years, the company will further deepen the cooperation with hospitals and expand the coverage and influence of service package products.

A：Many patients have shown acceptance of self-paid services in the promotion of hospitals, and 50% to 60% of patients are willing to pay for them personally. On the basis of not increasing the hospital's excessive resource investment, it actually helps the hospital to provide more service charges. This is in line with the guiding direction of the national medical reform policy, especially in the context of the price reform of equipment, consumables and drugs, the state encourages hospitals to increase service-oriented charges.

A：From the patient's perspective, the service ensures professional guidance of the rehabilitation process and avoids the need for additional medical investment and secondary admission due to lack of sufficient knowledge. From the doctor's perspective, the AI system can assist in the detailed guidance of discharged patients and improve medication compliance and follow-up rate. From the perspective of medical insurance, unplanned expenditure of medical insurance is reduced by avoiding unplanned readmission.

A：The company currently operates 14 hospitals, plans to expand to 28 by the end of the year, and has identified about 400 target hospitals nationwide. In the coming year, it will expand and operate the 93 largest hospitals in the country, and then build a demonstration effect of head hospital operation in more than 300 other hospitals, gradually radiating to various regions, covering more small Junior College hospitals, and realizing a three-step business development plan.

A：AI system can automatically generate detailed discharge guidance for patients and remind patients to check, review and take medicine in real time, significantly improving medication compliance (from 49% to 95%) and follow-up rate (from 39% to 67%), effectively solving the tracking management problem of patients after discharge, and reducing the recurrence rate and complication rate.

A：The competitive barriers of the image cloud platform mainly include the advantages of artificial intelligence technology for case quality control and image quality control, which can ensure the quality of image data accessed across hospitals and achieve rapid presentation, greatly reducing the waiting time of doctors. In addition, through the construction experience of the Anhui Provincial Image Cloud Project, the data cloud and interconnection of medical institutions across the province have been successfully realized, ensuring a high connectivity rate and data mutual recognition. This successful experience has been extended to Guangxi and other provinces and will Continue to replicate and expand nationwide.

A：The core technical barrier of image cloud mainly lies in solving the problem of unnecessary repeated examination, which requires AI technology to accurately judge whether an image examination report is qualified and whether the equipment is accurate, so as to help doctors decide whether to call the existing external hospital examination report when billing, so as to avoid repeated examination. At present, only xunfei has more accurate AI testing than a single imaging doctor and even more accurate than the consultation results of expert imaging doctors in the standardized testing of the health and health Committee and medical insurance.

A：The image cloud realizes cross-hospital access through second-level access technology, and doctors can view the image data of the outer hospital in real time on the workstation, which is much faster than the traditional way. In addition, the integration of his and px interfaces with many hospitals has been completed nationwide, ensuring that image data can be efficiently shared and called across provinces and hospitals.

A：The image cloud starts with the image cloud serving the National Health Insurance Bureau, and then expands to the pan-image cloud and pathology services, which can double the business volume of each province. AI capacity is then used to provide remote consultation and AI follow-up services to support the needs of doctors in the process. In the long run, the company will build the world's largest imaging data center and AI capability platform to provide online diagnostic services to the whole country and the world, including individual users can also obtain image cloud-based AI-assisted diagnosis and treatment.

A：For grassroots doctors, AI diagnosis and treatment assistants are used as general auxiliary diagnostic tools to help identify atypical symptoms and guide doctors to conduct standardized inspections, improve the level of diagnosis and treatment of grassroots doctors, and reduce misdiagnosis and irrational drug use. For expert doctors in secondary and tertiary hospitals, AI diagnosis and treatment assistants can improve the efficiency of diagnosis and treatment and broaden the perspective of general practice by improving the efficiency, such as automatically generating electronic medical records, pre-consultation and other means, so that doctors can focus more on communication with patients, and at the same time assist diagnosis and decision-making in Junior College fields, so as to realize virtual multidisciplinary consultation.