Simulink: Is there a difference between Keplerian propagator and point-mass high precision propagator?

In order to see the difference between the fast yet simple propagator and the high-precision propagator of Simulink, I created a simple model with both blocks.

All the parameters (start date and orbital elements) are the same for both blocks.

In the case of the high precision, I set the (Earth) gravity model to Point-mass.

It was my expectation that a point-mass model should produce a somewhat similar output as a Keplerian model. However, there are errors in the scale of 1e6-1e7 meters in position between the two. This is reflected, for example, in an ever-increasing value for the radius averaged over an orbit. In the point-mass Keplerian propagator, the average remains the same. In the point-mass high-precision propagator, the average increases indefinitely.

Could someone clarify what is the difference between the two blocks (when the high precision is set to point-mass gravity)?

Thank you