You need to learn how to tell the difference between a syntax highlighted Markdown Python code block and Python that was passed through the Code Interpreter tool, but there is a visual difference. Executed Python displays on a black background.

It can definitely make up reasoning (including code) for how it got to an answer. However o3 can run python actually. I tried uploading an image and it ran a bunch of scripts to crop and change the brightness in an attempt to get a clearer view of various features.