I'm impressed. There certainly are no "reasonable" solutions, where I am arbitrarily defining reasonable as numbers that humans can work or understand in their heads.

Isn't there a version of ChatGPT that connects to Wolfram Alpha? Did you try that one?