I can't tell if that is sarcasm. Of course you can use the same model to write tests. That's a different problem altogether, with a different series of prompts altogether!

When it comes to code review, though, it can be a good idea to pit multiple models against each other. I've relied on that trick from day 1.