Benzodiazepines have been widely used for their anxiolytic actions. However, the contribution of GABA A receptor subtypes to anxiolysis is still controversial. Studies with mutant mice harboring diazepam-insensitive α-subunits α1, α2, α3, or α5 have revealed that α2-containing GABA A receptors (α2-GABA A Rs) are required for diazepam-induced anxiolysis, with no evidence for an involvement of any other α-subunit, whereas TP003, described as a selective modulator of α3-containing GABA A receptors, was shown to be anxiolytic. Here, we describe a novel, systematic approach to evaluate the role of positive allosteric modulation of each of the four diazepam-sensitive α-subtypes in anxiety-related behavioral paradigms. By combining H to R point mutations in three out of the four diazepam-sensitive α-subunits in mice with a 129X1/SvJ background, diazepam becomes a subtype-specific modulator of the remaining non-mutated α-subtype. Modulation of α5-GABA A Rs, but not of α2-GABA A Rs, increased the time in the light side of the light-dark box as well as openarm exploration in the elevated plus maze. In contrast, modulation of α3-GABA A Rs decreased open-arm exploration, whereas modulation of α2-GABA A Rs increased time in the center in the open-field test. Modulation of any single α-subtype had no effect on stress-induced hyperthermia. Our results provide evidence that modulation of α5-GABA A Rs elicits anxiolytic-like actions, whereas our data do not provide evidence for an anxiolytic-like action of α3-GABA A Rs. Thus, α5-GABA A Rs may be suitable targets for novel anxiolytic drugs.