In recent strides within the field of Multimodal Large Language Models (MLLMs), while notable progress has been made in input-side multimodal comprehension, a notable void persists in the domain of multimodal content generation. Addressing this gap, a collaborative effort between Tencent AI Lab and The University of Sydney introduces GPT4Video in a new paper GPT4Video: A Unified Multimodal Large Language Model for lnstruction-Followed Understanding and Saf