If I understand correctly your question, you docked PRPc and Laminin and you don’t know which cluster is the most likely to be correct. Well, with so little insights into your run, I cannot help much.
All the clusters are sorted by HADDOCK score, averaged over the top 4 models of each cluster. All the statistics reported on the results page of the server are also averaged over the top 4 models of each cluster. The HADDOCK score is consistently performing great in blind challenges such as CASP and CAPRI. Therefore, the lower (= the better) the HADDOCK score, the higher is the chance that your cluster depicts the correct binding. But molecular docking is a complex problem and you should always try to confirm your model with extra experiments.