•Multi-objective optimization of gear pair parameters was conducted.•Main goal was reducing the transmission volume and power losses.•Five optimization variables were used; z1, b, m, x1, and x2.•Solutions have shown that the trade-off between volume and efficiency is obligatory.•Combination of the lower gear module, the lower face width, and the higher profile shift coefficients yield good results. Besides satisfying the essential strength requirements, gearbox design should ensure additional desirable properties in order to be competitive. For example, a gearbox should be efficient, durable, quiet, compact, and light. Nowadays, as a consequence of rising environmental concerns, high efficiency is a rather desirable feature. In this article, a genetic algorithm was used for conducting a multi-objective optimization of gear pair parameters with a goal of reducing the transmission volume and power losses. Gearing efficiency primarily depends on the normal load, sliding velocities, and the friction coefficient. Gearing efficiency was calculated analytically, using the approximate load distribution formulae and efficiency formulation developed by Schlenk. The resulting formula was included in the genetic algorithm as an objective. To verify it, results were compared to the ones obtained by other authors. Optimization variables consisted of the gear module, the face width, the pinion and wheel profile shift coefficients and the number of teeth of the pinion. Solutions have shown that the trade-off between volume and efficiency is obligatory and a combination of the lower gear module, the lower face width, the higher profile shift coefficients and the higher number of teeth of the pinion yield good results regarding both objectives.