For the technical quality assurance of breast cancer screening protocols several phantoms have been developed. Their dose sensitivity is a common topic often discussed in literature. The European protocol for the quality control of the physical and technical aspects of mammography screening suggests a contrast-detail phantom like the CDMAM phantom (Artinis Medical Systems, Elst, NL). The CDMAM 3.4 was tested with respect to its dose sensitivity and compared to other phantoms in a recent paper. The CDMAM 4.0 phantom provides other disc diameters and thicknesses adapted more closely to the image quality found in modern mammography systems. This motivates a comparison of the two generations using the same exposure parameters. We varied the time-current (mAs) within a range of clinically used values (40-140 mAs). All evaluations were done using automatic evaluation software provided by Artinis (for CDMAM 4.0) and the National Coordinating Centre for the Physics of Mammography, Guildford UK (CDMAM 3.4). We compared the relative dose sensitivity with respect to the different diameters and also computed the IQFinv parameter, which averages over the diameters as suggested in the manual for the phantom. The IQFinv parameter linearly depends on dose for both phantoms. The CDMAM 4.0 shows a more monotonous dependence on dose, the total variation of the threshold thicknesses as functions of the dose are significantly smaller than with the CDMAM 3.4. As the automatic evaluation shows rather different threshold thicknesses for the two phantoms, conversion factors for human to automatic readout have to be adapted.