The precision of the 26Al–26Mg system—one of the most widely used chronometers for constraining the relative timing of events in the early solar system—is presently limited by methods for the determination of 27Al/24Mg ratios, which have seen little improvement in the last decade. We present a novel method for the measurement of 27Al/24Mg ratios in unpurified sample solutions by multiple-collector inductively coupled plasma mass spectrometry. Because Al is monoisotopic we use a modified isotope dilution approach that employs a mixed spike containing isotopically enriched 25Mg and natural 27Al in an accurately known ratio. In order to determine the spike to sample ratio for Al, measurements of spiked aliquots are bracketed by unspiked aliquots, which negates the impact of elemental bias. Unlike conventional isotope dilution, samples do not require chromatographic separation prior to analysis, which both saves time and minimises the risk of contamination of other samples with spike (which is added immediately prior to analysis). Repeat measurements of the BHVO-2, BCR-2, and BIR-1 international rock standards, as well as a gravimetrically prepared Al–Mg reference solution, indicate that our method is both accurate and reproducible to 0.2%. This 4- to 10-fold improvement over previous methods translates directly to an equal gain in the resolution of the 26Al–26Mg chronometer. The approach presented here could, in principle, be applied to other monoisotopic elements such as the Mn–Cr system. Based on multiple measurements of a ∼2.7 gram piece of the Ivuna CI chondrite, we present a new estimate for the 27Al/24Mg ratio of this meteorite of 0.09781 ± 0.00029.