In this report, we delve into the capabilities of GPT-4, a large multimodal model that can seamlessly process both image and text inputs, and generate text outputs. As the applications for such models continue to expand in areas such as dialogue systems, text summarization, and machine translation, we aim to shed light on the intricacies of this model and the potential it holds. The report also discusses the recent developments in the field of multimodal models that have paved the way for GPT-4's creation [1–34].