Consider the structure $(\mathbb{Q},+,-,\times,0,1,<).$ Now certainly, by Lowenheim-Skolem, we can't have a set of sentences in first-order logic whose models are precisely that structure and any structure isomorphic to it. But what if we require the model to be countable? That is, if a countably infinite ordered field satisfies the theory of the rational ordered field, is it isomorphic to the rational ordered field?
Edit: Also, if the answer is no, can someone exhibit an ordered field not isomorphic to $\mathbb{Q}$ which satisfies its complete theory?
By compactness, $\mathbb{Q}$ has a non-Archimedean elementary extension $\mathcal{A}$. Of course $\mathcal{A}$ may be uncountable, but we can get around that: let $a\in\mathcal{A}$ be an infinite element, and apply downward Lowenheim-Skolem to get a countable $\mathcal{B}\preccurlyeq\mathcal{A}$ with $a\in \mathcal{B}$.
Since $\mathcal{B}\preccurlyeq\mathcal{A}\equiv\mathbb{Q}$ we have $\mathcal{B}\equiv\mathbb{Q}$. But by choice of $a$, $\mathcal{B}$ must be non-Archimedean - and so not isomorphic to $\mathbb{Q}$.
EDIT: Whoops, I was really wrong initially - here's a correct statement:
It turns out in fact that there's a very useful general principle available to us: if $T$ is a (complete, countable language) theory which interprets a non-$\aleph_0$-categorical theory, then $T$ itself is not $\aleph_0$-categorical. So we can prove non-categoricity just by finding "bad configurations." This is a consequence of the Ryll-Nardzewski theorem, that a (countable complete) first-order theory is $\aleph_0$-categorical iff it has finitely many $n$-types for each $n$. See Alex Kruckman's comment below.