We report on a series of alkyl- and alkoxy-substituted 1,4-dihydroquinoxaline-2,3-diones (QXs), prepared as a continuation of our structure-activity relationship (SAR) study of QXs as antagonists for the glycine site of the N-methyl-D-aspartate (NMDA) receptor. The in vitro potency of these antagonists was determined by displacement of the glycine site radioligand [3H]-5,7-dichlorokynurenic acid ([3H]DCKA) in rat brain cortical membranes. In general, methyl is a good replacement for chloro or bromo in the 6-position, and alkoxy-substituted QXs have lower potencies than alkyl- or halogen-substituted QXs. Ethyl-substituted QXs are generally less potent than methyl-substituted QXs, especially in the 6-position of 5,6,7-trisubstituted QXs. Fusion of a ring system at the 6,7-positions results in QXs with low potency. Several methyl-substituted QXs are potent glycine site antagonists that have surprisingly high in vivo activity in the maximal electroshock (MES) test in mice. Among these, 7-chloro-6-methyl-5-nitro QX (14g) (IC50 = 5 nM) and 7-bromo-6-methyl-5-nitro QX (14f) (IC50 = 9 nM) are comparable in potency to 6,7-dichloro-5-nitro QX (2) (ACEA 1021) as glycine site antagonists. QX 14g has an ED50 value of 1.2 mg/kg iv in the mouse MES assay. Interestingly, alkyl QXs with log P values of 0.5 or less tend to be more bioavailable than QXs with higher log P values. QX 14g has 440-fold selectivity for NMDA vs alpha-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid (AMPA) receptors, as determined electrophysiologically under steady-state conditions in oocytes expressing rat cerebral cortex poly(A)+ RNA. Overall, 14g was found to have the best combination of in vitro and in vivo potency of all the compounds tested in this and previous studies on the QX series.