User testing navigation with one or two groups

We want to test a set of five users to compare usability between our existing application navigation and our proposed next generation navigation. During the course of planning how we will do our testing. Currently our options are as follows:

  1. Test both navigations with the same group
  2. Test both navigations with a separate group for each

I was wondering if anyone had any advice on which has delivered a more optimal response. Option 1 has the possibility of the first test affecting the results of the second but I wanted to find out more opinions