As far as the implementation is concerned both should work fine. However, since we expect higher deformations in the elements due to plasticity. I would recommend using SOLID 186 whenever possible. As you might already know, higher order elements refer to the use of higher-order shape functions. The shape function represents assumed behavior for a given element and how well each element's shape function matches the true behavior directly affects the accuracy of the solution. As Peter said, these elements can represent curved edges and surfaces more accurately and are not as sensitive to element distortion. They have also been reported to predict highly accurate stress and give better results than linear elements in many cases even with fewer elements. In my experience, this difference between different ordered elements is less severe for shells in comparison to solids.
Here is a link to a really good article that I had bookmarked a while ago.
Now the additional aspect that affects the accuracy of the solution in addition to the element order is the number of integration points, as the manual here says a quadratic element has no more integration points than a linear element.
Stresses (and strains) are calculated at the integration points, not at the nodes, and the variation of stress across an element is determined by the number of integration points. Having one integration point for a particular stress yields a constant value through the thickness while having two (i.e., 2x2x2 in 3D) yields a linear variation. The number of nodes does not determine the number of integration points, and different element formulations may use different number of integration points.
(1) Uniform reduced integration uses one integration point. You will get a constant value for stress/strain within an element.
(2) Enhanced Strain or Simplified Enhanced Strain have 2x2x2=8 (in 3D) for both volumetric and deviatoric terms. You will get a linear variation of any stress quantity requested.