Dear Dr. Mathias
Thanks for your comments and sorry for the delayed reply. I tried both BGK and MRT, along with other boundary conditions – LocalBC, InterpBC and Zou-HeBC. All of them yield similar converged streamline/velocity profiles. So, I believe that this is how the physical solution should be.
Regards
Sthavishtha