The mechanisms behind the Amyloid-beta (A beta) peptide neurotoxicity in Alzheimer's disease are intensely studied and under debate. One suggested mechanism is that the peptides assemble in biological membranes to form beta-barrel shaped oligomeric pores that induce cell leakage. Direct detection of such putative assemblies and their exact oligomeric states is however complicated by a high level of heterogeneity. The theory consequently remains controversial, and the actual formation of pore structures is disputed. We herein overcome the heterogeneity problem by employing a native mass spectrometry approach and demonstrate that A beta(1-42) peptides form coclusters with membrane mimetic detergent micelles. The coclusters are gently ionized using nanoelectrospray and transferred into the mass spectrometer where the detergent molecules are stripped away using collisional activation. We show that A beta(1-42) indeed oligomerizes over time in the micellar environment, forming hexamers with collision cross sections in agreement with a general beta-barrel structure. We also show that such oligomers are maintained and even stabilized by addition of lipids. A beta(1-40) on the other hand form significantly lower amounts of oligomers, which are also of lower oligomeric state compared to A beta(1-42) oligomers. Our results thus support the oligomeric pore hypothesis as one important cell toxicity mechanism in Alzheimer's disease. The presented native mass spectrometry approach is a promising way to study such potentially very neurotoxic species and how they could be stabilized or destabilized by molecules of cellular or therapeutic relevance.