There are varieties of platonism, that, the mathematical properties, are ideals.
For some this is after axiomatics that the objects do have platonic reality,
or, that they don't.
Then, furthermore, some strong platonists, have the objects exist
even if "we" didn't exist to "axiomatize" them, that only purely formal
axioms that result in logical structure are "true axioms".
Then, some realists, are "not platonists", while, some are, "strong platonists",
that with respect to the operationalists, the "quantities are always in units"
with respect to "numbers: dimensionless quantities".
Examples of physics needing mathematics include the N-body problem,
wave equation, ..., functional freedom, parastatistics, .... I.e. that
mathematics owes physics strong enough mathematics for quite a
variety of concerns what must make for the derivations as in the applied.
(Physics writ large is already using an often very regular and usual stack
of derivations and mathematics, eg defined under formula.)
I totally agree that the threeness, of, of a quantity or a count or a number
or "anything in the equivalence class that the count of which is three",
is just that numbers are simply enough classes under values.
That that's a giant class while instead something like von Neumann ordinals
make for a model that also "an ordinal assignment" for a count, for that
"things with counts are in a cardinal, the equivalence class of sets with those",
sets, makes for that threeness is as direct as things with the number and
things with the count.
I.e., that threeness relates to fourness, counting separately each,
is as a larger space than even a cardinal, that Peano's counting is
about as drawn down as it can be while for example something like
the "Principia Mathematica to arrive at 1+1 = 2" has that the implicit
space exists for all the non-logical theories one model of numbers.
This is for that mathematics is a _part_ of (any) science, as is for pure
logic and then that numbers strongly platonically exist, that overall,
the theory, i.e. some ideal unified theory, is a science.
Surely I'll agree "there's a theory where it's immaterial whether I'm
a strong platonist or platonist where in terms of mostly reason or
even the empirical, apples are to apples and apples are to oranges,
that validity of arithmetic is so falsifiable is that invalidities of arithmetic
are falsifiable".
Here of course "falsifiable" means independent under concerns not "contradictable".
I.e., not necessarily contradictable where it's not false.
Formalism's key strengths include axiomatics and thusly the fundamental.
Fundamentalist formalism these days has ZF set theory the most explored
milieu for the descriptive, what resultingly most formalisms trace formalization.
Of course domains like real analysis or for example topology are as much
"independent in the descriptive", what that largely set theory and then for
type-theories-and-category-theories-sons-of-set-theory, if a strong platonist
is conscientious to bring that on down from utter fundamentalism,
that from pure logic in mathematics is usually numbers (or form, structure).
That the mathematical models for mathematical physics have a
physical interpretation (or the intended intepretation), is for that
there's the fundamental success of mathematics in physics for its
validities, that modeling the total or universal in mathematical objects,
needs mathematics of those kinds of infinities for physics.
Where for example real analysis with the differential for
the integral calculus for the applied is ubiquitous...,
in infinite limits.