The present study was focused on MM-GBSA and MM-PBSA model performance in activity evaluation of AMPA receptor pos. allosteric modulators with the dimer of ligand-binding domains of the GluA2 receptor.The largest data-set was obtained using FLIPR methodol., affinity was also measured by radio-labeled ligand substitution method and their statistical anal. was performed in R 3.0.2, plots were made with ggplot2 package.The results showed significance level of the number of rotatable bonds in MM-GBSA-based model is quite low, at the same time, the model created using MM-GBSA and the number of bound ligands demonstrates the significance of the latter variable and interestingly, a variable ′the number of rotatable bonds′ becomes significant in MM-PBSA-based model and allows one to increase R2 up to 0.70 but the sign of ′nrot′ coefficient is pos. which is not consistent with the theor. sign of entropy contribution.The author concluded that both GBSA and PBSA solvation models showed good performance for the described data-set and demonstrated that AMBER14 binding energy reasonably correlated with pEC50 for AMPA receptor pos. allosteric modulators.