Akaike information criterion and Cox regression

I am trying to compare my models, but I have some issues.

I use the equation
AIC = 2*k - 2 ln (Log Likelihood)

1. I don't know what to do with a -2, since the (-2 ) is already in the report from SPSS in a form of -2 Log Likelihood (red circle).

2. I have diferrent variables, continuous aren't of a question, but categorical, because some of them have more than two subgroups.
Therefore, I don't know what to use as k. I have two variables, but five or even seven subgrups in some cases.

For example - pictures 1 and 2.

Thank you.