The versatility of non-aqueous capillary electrophoresis (NACE) results mainly from the variety of physico-chemical properties of the different solvents. They provide solubility for a wide range of analytes, enable to control electrophoretic selectivity, but affect in some cases UV absorbance detection. The coupling of NACE to electrospray mass spectrometry (ESI-MS) allows to cope with the high UV cut-off of some CE relevant solvents (e.g., formamides). In this paper the pure organic solvents methanol, acetonitrile, dimethylsulfoxide, formamide, N-methylformamide and N,N-dimethylformamide are evaluated against water for the preparation of ammonium acetate electrolytes to separate the basic model substances 2-aminobenzimidazole, procaine, propranolol and quinine with NACE-MS. MS coupling is assisted with the sheath liquid water-isopropanol (1:4, v/v) with 0.1% formic acid. The goal of the paper is to assess the influence of the solvent on selectivity, separation speed, and peak efficiency for a given set of model compounds on a simple empirical basis. It should give the user an idea how the separation quality is changed when nothing but the running solvent is altered. The obtained efficiency results were discussed with respect to physico-chemical models described in literature (assuming longitudinal diffusion as the only source of band broadening), but no satisfying correlations with solvent properties could be traced. The feasibility of all six organic solvents for MS coupling was demonstrated and the influence of the separation solvent on the MS detection performance was compared. In the seven different solvents, the shortest run time was obtained with acetonitrile, the best peak resolution with the amphiprotic solvents (especially methanol) best peak efficiency with methanol and formamide, and the most sensitive ESI-MS detection with acetonitrile and methanol, but with only slight advantage to water.