Within our program of physical characterization of trans-Neptunian objects and centaurs, we predicted a stellar occultation by the centaur (54598) Bienor to occur on January 11, 2019, with good observability potential. We obtained high accuracy astrometric data to refine the prediction, resulting in a shadow path favorable for the Iberian Peninsula. This encouraged us to carry out an occultation observation campaign that resulted in five positive detections from four observing sites. This is the fourth centaur for which a multichord (more than two chords) stellar occultation has been observed so far, the other three being (2060) Chiron, (10199) Chariklo, and (95626) 2002 GZ 32 . From the analysis of the occultation chords, combined with the rotational light curve obtained shortly after the occultation, we determined that Bienor has an area-equivalent diameter of 150 ± 20 km. This diameter is ∼ 30 km smaller than the one obtained from thermal measurements. The position angle of the short axis of the best fitting ellipse obtained through the analysis of the stellar occultation does not match that of the spin axis derived from long-term photometric models. We also detected a strong irregularity in one of the minima of the rotational light curve that is present no matter the aspect angle at which the observations were done. We present different scenarios to reconcile the results from the different techniques. We did not detect secondary drops related to potential rings or satellites. Nonetheless, similar rings in size to that of Chariklo's cannot be discarded due to low data accuracy.