In theory, Eddy Current should find everything that DPI or MPI will, but there are certain limitations.
MPI is field direction sensitive, so you need to be certain that you have covered the component or area with crossing magnetic fields.
The same is applicable to Eddy Current. This will not easily detect a crack when the probe runs along the length of the crack, so once again, you need to double inspect at right angles to the first inspection to be certain of detecting the cracks.
Also when it comes to borderline defect dimensions, there are some limitations.
In a laboratory environment with experimental Eddy Current equipment, we have been sucessful in reliably detecting cracks of 0.020mm deep, whilst with a very sensitive penetrant or magnetic test, detection of smaller defects could be possible. Here however surface finish is critical for all three methods.
Eddy current can also detect defects under a thin non-conductive coating (like MT, but unlike PT).
DPI will is not limited to electricaly conductive material. Also, I have used direct current fluorescent MPI to locate flaws 1/4" deep. Of course that is limited to ferromagnetic material, while eddy current can be used to test other electricaly conductive materials.