model errors
You have no guarantee of anything on models. Your best bet is to simulate the test circuits in the device data sheet and see if you get the same results.
Some models like JFET are very inaccurate because of the limitations of device parameters in SPICE. I once simulated a source follower that had a simulated frequency response that went up to X-rays.
A rail to rail op amp model had a simulated output that exceeded the +supply voltage and went no nearer to the - rail than 2 V.