Question about extracting images from files with GPT4o


I am interested in a particular topic. I noticed how effective the new GPT-4 is at extracting images from PDF files when you upload them. However, it also shows me code where it uses PyPDF2 or other libraries to extract images.

I tried to find out how it performs this task when I upload a file. Is it using various Python libraries in the background or some kind of general intelligence? Because I want to make API calls where I can send a PDF file and instruct it to extract images, I am unsure if GPT-4 can do this directly, or if I will need to use some Python libraries. What is going on behind GPT-4?