In the absence of direct kinematic measurements, the mass of an accreting black hole is sometimes inferred from the X-ray spectral parameters of its accretion disk; specifically, from the temperature and normalization of a disk-blackbody model fit. Suitable corrections have to be introduced when the accretion rate approaches or exceeds the Eddington limit. We summarize phenomenological models that can explain the very high state, with apparently higher disk temperatures and lower inner-disk radii. Conversely, ultraluminous X-ray sources often contain cooler disks with large characteristic radii. We introduce another phenomenological model for this accretion state. We argue that a standard disk dominates the radiative output for radii larger than a characteristic transition radius , where is the accretion rate in Eddington units and RISCO is the innermost stable orbit. For RISCO < R < Rc, most of the accretion power is released via non-thermal processes. We predict the location of such sources in a luminosity–temperature plot. We conclude that BHs with masses ∼50–100M⊙ accreting at may explain the X-ray properties of many ULXs.