I need to calculate some UV-vis spectra, where most interesting transitions are due to intermolecular charge transfer. Size of a problem might scale up to hundred of second and third row elements atoms, hence CCSD(T), which is as I know the best way, isn't applicable, DFT is preferable, but I understand its drawbacks.Currently thinking about CAM-B3LYP, for example, but have some doubts. Will be grateful for any good benchmark references.