The field of video understanding and multimodal large language models is rapidly evolving, with a focus on developing models that can reason and understand complex video content. Recent research has highlighted the importance of evaluating and advancing the reasoning capabilities of video models, with a particular emphasis on cross-video reasoning, think-in-video reasoning, and generative visual reasoning. Noteworthy papers in this area include CrossVid, which introduces a comprehensive benchmark for evaluating cross-video reasoning in multimodal large language models, and TiViBench, which proposes a hierarchical benchmark for evaluating the reasoning capabilities of image-to-video generation models. Additionally, Gen-ViRe and V-ReasonBench provide frameworks for assessing video models' reasoning abilities, while VR-Bench explores the paradigm of reasoning via video generation. These advancements have the potential to significantly improve the performance and reliability of video models, enabling them to better understand and generate complex video content.