On six different species of benthic foraminifera covering various microhabitats and recovered from the Northern Atlantic Ocean, we tested the potential of the Mg/Ca and Sr/Ca ratios as proxies for paleoceanography. We performed analysis using two instruments (ICP-OES and ICP-MS) and compared results obtained from living and dead specimens. Our results are in good agreement with previous published calibrations for Hoeglundina elegans, Uvigerina mediterranea, U. peregrina, Melonis barleeanum, and Globobulimina spp. However, we observed a strong variability between living and dead specimens, and between both instrumental approaches. We discuss the impact of the cleaning procedure, as well as the natural variability between samples recovered at different depths inside the sediment. No specific trend can be deciphered from our dataset, but we observed that species from the Uvigerina genus presented the lowest external reproducibility and the best agreement between living and dead specimens. We highlight that both species should not be mixed for analysis, since U. mediterranea presents lower values and a reduced range of variability compared to U. peregrina. We explored the temperature and the Δ[CO32−] as potential controls on the variability of both ratios from U. peregrina and showed that neither of these two parameters can be discarded.