Aims. We derive constraints on energy release, transport and conversion processes in solar flares based on a detailed characterization of the physical parameters of both the thermal plasma and the accelerated nonthermal electrons based on X-ray observations. In particular, we address the questions of whether the energy required to heat the thermal plasma can be supplied by nonthermal particles, and how the energetics derived from X-rays compare to the total bolometric radiated energy. Methods. Time series of spectral fits and images for 24 flares ranging from GOES class C3.4 to X17.2 were obtained using RHESSI hard X-ray observations. This has been supplemented by GOES soft X-ray fluxes. In our companion Paper I, we have used this data set to obtain the basic physical parameters for the thermal plasma (using the isothermal approximation) and the injected energetic electrons (assuming the thick-target model). Here, we used this data set to derive the flare energetics, including thermal energy, radiative and conductive energy loss, gravitational and flow energy of the plasma, and kinetic energy of the injected electrons. We studied how the thermal energies compare to the energy in nonthermal electrons, and how the various energetics and energy partition depend on flare importance. Results. All flare energetics show a good to excellent correlation with the peak GOES flux. The gravitational energy of the evaporated plasma and the kinetic energy of plasma flows can be neglected in the discussion of flare energetics. The radiative energy losses are comparable to the maximum thermal energy, while the conductive losses are considerably higher than the maximum thermal energy, especially in weaker flares. The total heating requirement of the hot plasma amounts to ≈50% of the total bolometric energy loss, with the conductive losses as a major contribution. The nonthermal energy input by energetic electrons is not sufficient to account for the total heating requirements of the hot plasma or for the bolometric losses, in particular in weak flares. Conclusions. Our results support the standard model of solar flares, with the following modifications. (1) Heating the hot thermal plasma and supplying the bolometric radiated energy requires an additional non-beam heating mechanism. (2) Strong conductive losses are a necessary additional energy transport process that transfers the energy released in the corona to the lower (and denser) atmospheric layers, where the bulk of the released energy is efficiently radiated away at longer wavelengths (EUV, UV, and white light) by cooler material.