Molecular complexes and clusters provide bridges between molecular and solid states physics. Containing tens to few thousands of atoms, such systems can hardly be approached via traditional ab initio wavefunction based methods at the moment. Density functional theory (DFT) and density functional based tight binding methods (DFTB) have strongly developed with respect to computational efficiency to cover this size range. However both DFT and currently implemented DFTB face difficulties to describe realistically and accurately the typical interactions met in molecular clusters, in particular long range interactions such as Coulomb interactions between distant charge fluctuations, charge resonance in ionic clusters, and van der Waals interactions. The present article aims at providing an overview of how extensions of DFTB can circumvent some of the above deficiencies and turn out to be realistic and efficient tools to investigate the properties of molecular clusters and complexes, focusing on structural, electronic, energetic, and spectroscopic properties.