Practical value for stability factor

In design of a 10-GHz amplifier, what value for stability factor (k) is satisfactory? A value much greater than 1 is achieved in the expense of reduced gain, while a value near 1 is not reliable.

The stability factor problem has been discussed here in details.If you search , I'm sure you'll find very useful infornation about it.

Like in theory:

K>1
B1>0

This is obvious! Both K=1.01 and K=5 are greater than 1! However, the first is sensitive to process variation, and the later results in reduced gain. I need a practical solution.

Put in your simulation gain pplot and stability and look at both of them...
in general if Kgreater than 1.5 its pretty good enough ...

Do a Monte Carlo simulation and observe the variance of K stability factor by run..