Comment by floodfx

Why are completion tokens more with image prompts yet the text output was about the same?