Kinematic hardening question

[kostast88]

I performed a benchmark test of a simple elastic beamcolumn with a steel01 material, zero length element on it's base with kinematic hardening. I created the exact same model on sap2000. The results I got for elastic static and time history analyses without damping are exactly the same.
In sap2000 I used NLlink mltilinear plastic, giving coordinates on points of the backbone curve,instead of slope like in opensees.

For nonlinear time history (again without damping) though, there was a slight difference. When I checked the zero length moment-rotation inelastic cycles,I noticed that the hardening part of the line, had different slope in the two programs(!).

Since kinematic hardening in both programs does not take input parameters, I want to ask whether anyone has any knowledge of differences between these two programs on this subject, or an idea why this might be happening.