A comparative analysis of models from the IPCC archive would begin by extracting information from the IPCC Model Documentation Table. Let's take as an example Guilyardi (2006), a key paper by my friend Eric Guilyardi comparing characteristics of ENSO across all the models in the IPCC AR4 archive. Table 2 from that paper, reproduced here:

shows ENSO characteristics from many IPCC models. If you asked yourself, what's the dependency of ENSO on model resolution? you'd have to refer back to the IPCC documentation and retrieve this information, model by model, quite laboriously. This information is summarized in Table 1 from that paper, shown here:

Could a standard description of grids make it easy to create Table 1? Here is a beginning attempt at standardizing this information, based on NMM Schema.