Dear Molpro community, There is a large discrepancy between the analytical and numerical gradients I obtain for the Sapporo-DKH3-DZP-2012 basis set. Does Molpro support the use of analytical gradients with relativistic corrections? Thank you, Ryan Zaari