[marcosdumay]

Oh, that's right.

But magnetization is famously path-dependent, so now I'm tending to believe the force will be closer to 1/r^7 when moving on the direction of the magnet, but 1/r^4 when moving away. And not exactly any of those functions.

[planede]

Do you mean that the ferromagnetic material gets magnetized and retains some magnetization while moving away?

[marcosdumay]

Yes.