The brightest X-ray source in M82 has been thought to be an intermediatemass black hole (10 2−4 solar masses, M ⊙ ) because of its extremely high luminosity and variability characteristics 1−6 , although some models suggest that its mass may be only ∼ 20 M ⊙ 3,7 . The previous mass estimates are based on scaling relations which use low-frequency characteristic timescales which have large intrinsic uncertainties 8,9 . In stellar-mass black holes we know that the high frequency quasi-periodic oscillations that occur in a 3:2 ratio (100-450 Hz) are stable and scale inversely with black hole mass with a reasonably small dispersion 10−15 . The discovery of such stable oscillations thus potentially offers an alternative and less ambiguous mass determination for intermediate-mass black holes, but has hitherto not been realized. Here, we report stable, twin-peak (3:2 frequency ratio) X-ray quasi-periodic oscillations from M82 X-1 at the frequencies of 3.32±0.06Hz and 5.07±0.06 Hz. Assuming that we can scale the stellar-mass relationship, we estimate its black hole mass to be 428±105 M ⊙ . In addition, we can estimate the mass using the relativistic precession model, from which we get a value ofOscillations arising from general relativistic effects should scale inversely with the black hole mass if they arise from orbital motion near the innermost stable circular orbit in the We detected two power spectral peaks at 3.32±0.06 Hz (coherence, Q = centroid frequency (ν)/width(∆ν) > 27) and 5.07±0.06 Hz (Q > 40) consistent with a 3:2 frequency ratio ( Fig. 1a, b). The combined statistical significance of the detection is greater than 4.7σ (see Methods for details).The proportional counter array's field of view (1 • ×1 • ) of M82 includes a number of accreting X-ray sources in addition to M82 X-1 18 . The remarkable stability of the two quasiperiodic oscillations on timescales of a few years (Movies 1 & 2), their 3:2 frequency ratio and their high oscillation luminosities strongly suggest they are not low-frequency quasi-periodic oscillations from a contaminating stellar-mass black hole (see Methods for details). Also a pulsar origin is very unlikely for several reasons. First, a pulsar signal would be much more coherent than that of the observed quasi-periodic oscillations, which clearly have a finite width. Second, based on the observed high quasi-periodic oscillation luminosities it is extremely implausible that they originate from a pulsar (see Methods for details). Finally, it would be highly coincidental to have two pulsars in the same field of view with spins in the 3:2 ratio. Also, based on the average power spectra of the background sky and a sample of accreting super-massive black holes monitored by the proportional counter array in the same epoch as M82, we rule out an instrumental origin for these oscillations 3 & 4). This leaves M82 X-1, persistently the brightest source in the field of view, as the most likely source associated with the 3:2 ratio quasi-periodic oscillation pair.We estimated M82 X-1's black hole ...