SenseTime has open-sourced its multi-task vision model SenseNova-Vision-7B-MoT, aiming to provide a foundation for visual understanding and GUI intelligent agents development. The model integrates core vision tasks—including object detection, OCR, depth estimation, normal estimation, image segmentation, and multi-view processing—into a single 7B parameter model, and supports defining new vision task variants through natural language. SenseTime adopted a fully open-source strategy, allowing developers to obtain model weights, access a subset of the SenseNova-Vision corpus containing 50 million samples, and a complete reproduction toolkit for processing remaining public data. Industry experts note the release provides a more scalable unified framework for visual understanding tasks. Technical details, code repositories, and demonstration spaces are available on Hugging Face and GitHub.