Context.Molecular maser lines are signposts of high-mass star formation, probing the excitation and kinematics of very compact regions in the close environment of young stellar objects and providing useful targets for trigonometric parallax measurements.Aims.Only a few NH3(9,6) masers are known so far, and their origin is still poorly understood. Here we aim to find new NH3(9,6) masers to provide a better observational basis for studying their role in high-mass star-forming regions.Methods.We carried out NH3(9,6) observations toward Cepheus A and G34.26+0.15 with the Effelsberg 100-meter telescope (beam size 49′′) and theKarl G. JanskyVery Large Array (JVLA; beam size about 1′′.2).Results.We discovered new NH3(9,6) masers in Cep A and G34.26+0.25, which increases the number of known high-mass star-forming regions hosting NH3(9,6) masers from five to seven. Long-term monitoring (20 months) at Effelsberg shows that the intensity of the (9,6) maser in G34.26+0.25 is decreasing, while the Cep A maser remains stable. Compared to the Effelsberg data and assuming linear variations between the epochs of observation, the JVLA data indicate no missing flux. This suggests that the NH3(9,6) emission arises from single compact emission regions that are not resolved by the interferometric measurements. As JVLA imaging shows, the NH3(9,6) emission in Cep A originates from a sub-arcsecond-sized region, slightly to the west (0′′.28 ± 0′′.10) of the peak position of the 1.36 cm continuum object, HW2. In G34.26+0.25, three NH3(9,6) maser spots are observed: one is close to the head of the cometary ultracompact H IIregion C, and the other two are emitted from a compact region to the west of the hypercompact H IIregion A.Conclusions.The newly found (9,6) masers appear to be related to outflows. The higher angular resolution of JVLA and very long baseline interferometry observations are needed to provide more accurate positions and constraints for pumping scenarios.