Make OpenAI Vision API Match GPT4 Vision

Based on this, I understand it’s possible for the files ingested by GPT4 to contain images?
I have a knowledge base pdf containing a lot of screenshots I’d need to use.