Assuming a sufficient low frequency, e.g. a few 10 MHz for the shown geometry, the analysis can be restricted to AC magnetic, not considering the coupling between electric and magnetic field. If you also ignore skin effect and eddy currents, the problem reduces to an magnetostatic one. Then the coupling can be calculated by integrating Biot-Savarts law. This is almost easy for circular loops of axisymmetric arrangement and less convenient for spiral coils.
EM solvers also models skin and proximity effect as well as eddy currents. Coils of extreme geometries, where the current has to be carried by a very small part of the conductor area give possibly inaccurate results at high Q. But apart from this point, there's no problem with large or small structures. Small structure analysis are more accurate because skin depth spans a larger part of the structure.