Benchmarking how well LLMs can play FizzBuzz

I got this silly idea for a benchmark and decided to test how well models do. Repo here: https://github.com/venkatasg/fizzbuzz-bench