Ultrasound imaging plays a crucial role in biomedical diagnostics due to its non-invasive, real-time, and cost-effective nature. It is widely used in the examination of various organs and the detection of tumors. However, the accurate classification and segmentation of ultrasound images, especially for different organs and tumors, remain challenging tasks. Existing methods often struggle to achieve high performance across diverse anatomical regions and pathologies. From a technical perspective, the development of a universal model capable of handling multiple tasks in ultrasound image analysis is highly desirable. This would require advanced machine learning and image processing techniques to extract meaningful features and make accurate predictions. Such a model could potentially overcome the limitations of current approaches, which are often designed for specific organs or tasks and lack generalization ability. The envisioned impact of this challenge is twofold. Technically, it offers a great platform to test general model paradigms, spurring the research community to develop novel algorithms and models for better ultrasound image analysis. This may lead to discoveries like new feature extraction methods and optimized strategies, vital for validating models. Biomedically, the challenge outcomes will push models towards generalization. A precise universal model can aid clinicians in diagnosis, enabling earlier and more effective treatments, reducing invasive procedures, and highlighting the need for model evolution in medicine. The competition datasets consist of two collections: a public dataset bundle and a dataset bundle provided by 3 partner hospitals. The public dataset bundle includes 7 organs, with a total of 6740 ultrasound images, and contains both segmentation and classification tasks. The dataset bundle from the partner hospital also includes 7 organs, with 5262 images, and similarly includes both segmentation and classification tasks.