I have tested out my BGR circuit with ideal resistors and now I am trying to put in the real resistors, one would expect the real resistors to give the same results if the value used is almost the same as the value of the ideal resistors, however I was observing some major changes. To test the problem out I simply connected the real resistor to a voltage supply of 54mV with a value around 11k, in this the current value should be around 5uA, however I am getting a value around 15uA. Can anybody explain this discrepancy.

What is the temperature of the simulation (it is probably 25C if unspecified). I tried the same simulation at 25C and I get a different result With 54mV i get this:

ngspice 204 -> print i(v1)
i(v1) = -3.25486e-05

You see also the resistance changes a lot with voltage . The graph in the picture shows the resistance (VRES/I(V1) as a function of VRES. For your circuit please use res_xhigh_po and don't use the smallest geometry, since this makes the resistance very variable on biasing conditions. Use res_xhigh_po_1p41