2M1510 is an approximately 45-million-year-old triple system of brown dwarfs, two of which form a close binary in a 20-day orbit and have almost the same mass (3.82% and 3.75% of the mass of the Sun, respectively). Their physical parameters are in good agreement with evolutionary models except for luminosity, suggesting that we might be underestimating the masses of brown dwarfs and massive exoplanets by about 30%. Mass, radius and age are three of the most fundamental parameters for celestial objects, enabling insight into the evolution and internal physics of stars, brown dwarfs and planets. Brown dwarfs are hydrogen-rich objects that are unable to sustain core fusion reactions but are supported against collapse by electron degeneracy pressure(1). As they age, brown dwarfs cool, reducing their radius and luminosity. Young exoplanets follow a similar behaviour. Brown dwarf evolutionary models are relied upon to infer the masses, radii and ages of young brown dwarfs(2,3). Similar models are also used to infer the mass and radius of directly imaged exoplanets(4). Unfortunately, only sparse empirical mass, radius and age measurements are currently available, and so the models remain mostly unvalidated. Double-line eclipsing binaries provide the most direct route towards the absolute determination of the masses and radii of stars(5-7). Here we report the discovery by SPECULOOS (Search for habitable Planets EClipsing ULtra-cOOl Stars) of the 2M1510A triple system, consisting of a nearby, eclipsing, double-line brown dwarf binary and a widely separated tertiary brown dwarf companion. We find that the system is a member of Argus, a 45 +/- 5 million-year-old moving group(8,9). The system's age matches those of currently known directly imaged exoplanets so 2M1510A provides an opportunity to benchmark evolutionary models of brown dwarfs and young planets. We find that widely used evolutionary models(3) do reproduce the mass, radius and age of the binary components remarkably well, but overestimate their luminosity by up to 0.65 magnitudes, which could result in underestimations of 20% to 35% of photometric masses for directly imaged exoplanets and young-field brown dwarfs.