This is a fairly common problem with memory controller models. The model creator cut some corners and used one [Model Selector] keyword in the IBIS model for all the DDR signal pins. In order for the HyperLynx DDR wizard to benefit from, and use the model selector data, the different groups of signals need to reference different model selector labels. Here is a technote on SupportNet that describes the IBIS syntax and how to change the IBIS file to make it more useful.
The title of the technote is about receiver thresholds, but the solution applies to your situation.
Thank you Weston,
After segregating all the DDR signal groups these issues have been cleaned up. Looks like I was going to have to go thorugh this anyways in order to address the receiver thresholds that were referenced in the tech note you pointed me to, the vendors IBIS model falls short there as well.
I'm glad that the technote helped. These are fairly common problems in DDR controller IBIS models. It would help the industry if you reported back to the IBIS file provider how to make their models better.