You’ll have to use 3.1. Because of Rosetta’s documentation problems, we decided not to release applications that were insufficiently documented in 3.2; the RNA suite is the most significant application I’m aware of that was “unreleased” in 3.1. If you’d like to see it again in 3.3, write to its corresponding authors and encourage them to get it back up to speed.
DNA is typically not thought of forming much in the way of complicated structures by itself – at least not in comparison to RNA. Is there some indication that your DNA molecule is doing something other than forming a standard double helix, or being an unstructured polymer? Are you trying to model the DNA by itself, or in complex with something else (like a binding protein)?
If you do expect non-trivial structure for just the DNA, you may be able to repurpose the RNA structure prediction machinery. I don’t have any experience with nucleic acid modeling myself, but we just released version 3.4 which has a number of new applications specifically for RNA modeling, so looking though the documentation and demos there should give you a hint of what’s possible with Rosetta.