When all electrons are paired (and the spin fields cancel), all that is left is that Lorentz force and you get diamagnetism. Otherwise, you generally get paramagnetism (attractive because of the net dipole created by net spin).
See the Stern Gerlach experiment: https://en.wikipedia.org/wiki/Stern%E2%80%93Gerlach_experime...
That is the 1800's classical explanation of electric charge, which is not consistent with experiments, nor does it line up with magnetism, which is inherently quantum.
1. Even in your classical description, what you just described (following field lines) does not require there to be a gradient. If there is a gradient, yes, the forces will follow it. But there is still a force in a static infinite applied field, and potential energies still change as things move.
2. Diamagnetism is a quantum property, not a classical one. It exists due to the polarity of spins in an electron pair (although frankly, this explanation is weak and magnetism is one of the less understood subjects in physics). Even if a macroscopic field has a gradient, it will look like a static field at the scale of an atom, at which scale diamagnetic forces exist.
1. Yes, a gradient is required. If the energy of the diamagnetic material does not change, there cannot be a force, as this will violate conservation of energy. You are proposing a perpetual motion machine of the first kind.
2. The quantum nature of diamagnetism doesn't matter in the slightest to the argument being made, so why are you bringing this up?
Potential energy changes when moving parallel to any field force, regardless of whether that field is static or has some gradient.
Diamagnetism works by inducing dipoles in the medium, and a dipole placed in a constant field experiences no net force. It may experience a torque if it is not aligned with the field, but in the case of diamagnetism the induced dipoles are naturally aligned, and so there is neither force nor torque.