Many geomorphologic applications, notably glacier chronologies, require improvements in both the precision and the accuracy of the cosmogenic dating tool. Of particular importance is the need to better constrain the spatial variability of the cosmogenic nuclides production rates at high elevation and low latitudes. One strategy that can be adopted for this is to couple absolute calibrations, from independently dated surfaces, with cross-calibration studies, performed by measuring several cosmogenic nuclides in the same rock. In the present study, we report the highest-elevation (>4800 m) cross-calibration published to date, comprising measurements of cosmogenic 3He and 10Be in cogenetic pyroxene and quartz. The samples were collected from six dacitic moraine boulders, exposed from 32 to 65 ka at 4820 m on the flanks of the Uturuncu volcano (22° S, 67° W), Southern Lipez (Bolivia). The samples yield a remarkably tight cluster of 3He–10Be production ratios, with a weighted mean of 33.3 ± 0.9 (1σ). This production ratio is undistinguishable, within uncertainties, from the 3He–10Be production ratio of 32.3 ± 0.9 determined in the same mineral pair at low elevation (1333 m) by Amidon et al. (2009). These results agree at the 1σ level and suggest that any hypothetical increase of the 3He–10Be production ratio in pyroxene and quartz is likely to be lower than 5% over this elevation range (1000–5000 m). Moreover, the production ratio is almost insensitive to the Li content of the pyroxene (20 to 50 ppm Li), suggesting that the cosmogenic thermal neutron production of 3He is very low in this setting. The high-elevation 3He–10Be production ratio is used in combination with a local determination of the 3He production rate in the high Central Altiplano (3800 m) (Blard et al., 2013) to establish a local 10Be production rate of 30.0 ± 1.4at g-1 yr-1 at 3800 m and 20° S. After scaling to sea-level high latitude with the time-dependent Lal/Stone model, this yields a 10Be production rate in quartz of 3.63 ± 0.17at g-1 yr-1 Importantly, this rate can be used for high-precision geomorphologic dating, for example for determining glacial chronologies ( 1σ < 4%) through 10Be dating of moraines in the high Tropical Andes. Any inaccuracy attached to the scaling model is canceled out when the calibration site is located close to the dated object. The same experiment was also undertaken in pyroxene and plagioclase from two andesitic boulders from the Tunupa volcano moraines, exposed at 3800 m and 4200 m. A 3He–10Be production ratio of 35.9 ± 1.3 (1σ) is obtained for this mineral pair, indicating that the 10Be production rate in plagioclase is about 8% lower than in quartz.