Optimum conditions leading to a maximum sensitivity for both magnetodiodes and magnetoconcentration sensors are analyzed. These devices can operate in the accumulation as well as in the depletion regime of charge carriers, but the carrier accumulation is shown to yield the highest sensitivity if the sensor is provided with asymmetrical surface recombination rates and submitted to strong electric fields. Moreover, the sensor thickness must be carefully selected according to the proposed application: it has to be comparable with the diffusion length when detecting very small magnetic fields, but much thinner when measuring moderate fields or small fluctuations. Relative sensitivities exceeding several thousands of percent per tesla can then be achieved.