Multilevel SEM with complex sampling

I would say yes and no to this question. OpenMx supports ordinal data, complex sampling survey data, and user-specified fit functions. Ordinal data is supported by use of thresholds: http://openmx.psyc.virginia.edu/docs/openmx/latest/Ordinal_Path.html. Complex sampling survey data is supported by integration of sampling weights: http://openmx.psyc.virginia.edu/thread/3661. Some multilevel modeling is possible by use by "Wide" data formats. Latent Growth Curves are a typical example of using wide data to construct a multilevel model in a SEM program.

It would certainly be possible to put together an example using sampling weights, ordinal data, and a limited amount of multilevel structure. However, OpenMx does not currently have built-in features that make multilevel modeling particularly convenient. These are currently under active development by the OpenMx team.

I hope this answers your question.