<p>Video captioning (VC) is the task of automatically generating textual descriptions for video content, playing a crucial role in multimedia understanding and accessibility. It combines computer vision and natural language processing to interpret and describe visual scenes, making it valuable for applications such as content retrieval, video summarization, and assistive technologies. Video captioning has seen significant advancements with the rise of state-of-the-art (SoTA) models like ChatGPT, Claude, and Gemini. However, there is no clear comparison of these models in the context of VC, making it difficult for researchers and developers to choose the right one for their needs. This paper provides a structured comparison of these models, analyzing their strengths, weaknesses, and best use cases. By examining their performance, efficiency, and adaptability to different types of video content, we offer a practical guide to help users decide when to use each model. Our results show that Claude leads on precision-oriented metrics (BLEU and CIDEr), Gemini performs best on recall-oriented and semantic metrics (ROUGE and BERTScore), and GPT provides balanced performance across categories, with Gemini being the most cost-effective option for high-throughput VC tasks.</p>