Recent reports claim that modern light-emitting diode (LED) curing units improve curing efficiency by increasing the units' irradiance. In this context also, short polymerisation times up to 5 s are proposed. The aim of this study was to examine whether there are differences in the curing efficiency of modern LED curing units by assessing their effect on two different composite materials and by varying the irradiation time. A nano- and a micro-hybrid resin-based composite (RBC) were polymerised for 5, 10 and 20 s with three commercial and a Prototype LED unit (Elipar™ S10). Cylindrical specimens (6 mm in depth, 4 mm in diameter) were prepared in three increments, each 2-mm thick, and were consecutively cured. Degree of cure was measured for 20 min in real time at the bottom of the samples, starting with the photoinitiation. The micro-mechanical properties (modulus of elasticity, E and Vickers hardness, HV) were measured as a function of depth, in 100-μm steps, on the above described samples stored in distilled water for 24 h at 37°C. Data were analysed with multivariate ANOVA followed by Tukey's test, t test and partial eta-squared statistics. In descending order of the strength of their effect, the type of RBC, depth, polymerisation time and curing unit were significant factors affecting the micro-mechanical parameters (p < 0.05). The degree of cure at 6-mm depth was less but significantly influenced by the curing unit and curing time and was independent from the type of RBC. A 5-s irradiation time is not recommended for these units. Whereas a 5-s irradiation is acceptable at the sample's surface, a minimum of 20 s of irradiation is necessary for an adequate polymerisation 2 mm beyond the surface.