This isn't a good test for any model since LLMs can't math (even though frontier models can sometimes correctly simulate mathing), which is why one would always

This isn't a good test for any model since LLMs can't math (even though frontier models can sometimes correctly simulate mathing), which is why one would always use a tool for this.