On a general note, the waist is not a radial distance, but the location in space where the beam has the smallest width (there can be several waists, because of lenses). There is a well detailed explanation is the documentation for the 'Paraxial Gaussian Beam'.

It explains the two values for point 2. The documentation also shows the other columns:

The 'Waist' column is the radial width at the waist, which is the same for your surfaces 0 to 3 (propagation in air I guess). Thus width at the waist and the width at the surface ('Size' column) should match only when the waist is located at the surface (which is the case for your surfaces 0 and 1, probably co-located).

Regarding point 3, the X and Y width define the width of the array where the beam is stored. As Gaussian beams extend beyond the width at the waist, you want to have the array size larger than the full-width at the waist to store most of the beam. To quote the documentation, 'X- Y- Width The initial width in lens units of the region represented by the array.'

You really should refer to the documentation, it contains more details.