It's not math at all, it's words being generated by an LLM that cannot do basic arithmetic to save it's life. You don't run an equation through 2 other AIs to validate it, you test it against relevant constraints. This is basic validation you are capable of doing yourself.
On top of that - if you're reluctant to post the results online then why would you give it to a random individual? You can post your "findings" now and absolve yourself of any doubt, or take my word for it. AI generates bunk. The chances of it being right are practically lower than Siri or Clippy.