TMTPOST -- Alibaba Cloud's Tongyi Qianwen unveiled the industry's first open-source multimodal inference model, QVQ-72B-Preview. The QVQ model demonstrates exceptional visual understanding and reasoning capabilities, particularly excelling in solving complex reasoning problems in fields such as mathematics, physics, and science. Developers can now directly experience the model on platforms like MOTA Community and HuggingFace.