Core Insights - The Ctrl-World model, developed by a team from Tsinghua University led by Jianyu Chen and in collaboration with Stanford's Chelsea Finn, achieved the top ranking in the global evaluation of embodied intelligence by WorldArena [1] - The model excelled in four key dimensions: subject consistency, trajectory accuracy, depth accuracy, and strategy evaluation consistency [1] - In video generation capabilities, Ctrl-World ranked second globally, surpassing leading models from Google and NVIDIA, only behind Alibaba's Wan 2.6 [1] - Ctrl-World is recognized as a top-tier model in both "video generation quality" and "embodied tasks," indicating its dual strength in producing realistic videos and practical usability [1]