For plates and geogrids the stiffness properties must be specified in a stiffness per unit width. A common mistake is to assume that for axisymmetric problems the stiffness must be specified per radian, but this is not correct. For example, if a horizontal plate is defined on ground level one radian represents a pie slice that becomes wider when further away from the axis of symmetry. If the stiffness would be defined per radian this would mean that the stiffness of the pie slice remains constant when going further away from the axis of symmetry. However, further away from the axis of symmetry the pie slice is wider and thus has a lot more plate material. The only way to keep the stiffness constant when the plate becomes wider is either decrease the Young's modulus of the plate or decrease the thickness of the plate. Both cases are not practical: a circular plate that has been assigned one material set is expected to be of one and the same material and have constant thickness. This means the stiffness must be defined dependent on the width of the pie slice, that is per meter out-of plane.
For node-to-node anchors it is different since the anchors are not continuous in out-of-plane direction but line elements placed likes spokes in a wheel. The distance between the node-to-node anchors indeed increases when moving further away from the axis of symmetry, and according to the analogy of the wheel the number of node-to-node anchors (spokes) is constant per slice of the wheel. This means that anchor properties are per radian, and not per meter out-of-plane. In Plaxis the input for node-to-node anchors is split in two parts; a stiffness per anchor and a so-called spacing that indicates the distance out-of-plane in between the anchors. For an axisymmetric case this spacing for node-to-node anchors is in radians. Note that the spacing for fixed-end anchors is still per meter out of plane in an axisymmetric model.