> Mathematical proofs don't use models.

Maybe the person you replied to was taking a model theoretic perspective?

https://plato.stanford.edu/entries/modeltheory-fo/