I don't think it can be generalized like that.

Theoretically, Rust's memory ownership/aliasing rules are stricter and more granular than C's restrict, so some pointer-heavy code could optimize better. Rust is very good at inlining by default. Rust makes it easy to use stack-allocated structures. But C can do the same, it's just a matter of effort.

Both languages are low level enough that you can use them as a portable assembly and endlessly tweak them to one-up the other. If some C code doesn't optimize well you can write a more contorted code that will.