jats:titleAbstract</jats:title>jats:pCrystallization temperatures of primitive olivine crystals have been widely used as both a proxy for, or an intermediate step in calculating, mantle temperatures. The olivine‐spinel aluminum‐exchange thermometer has been applied to samples from mid‐ocean ridges and large igneous provinces, yielding considerable variability in olivine crystallization temperatures. We supplement the existing data with new crystallization temperature estimates for Hawaii, between 1282 ± 21 and 1375 ± 19°C. Magmatic temperatures may be linked to mantle temperatures if the thermal changes during melting can be quantified. The magnitude of this temperature change depends on melt fraction, itself controlled by mantle temperature, mantle composition and lithosphere thickness. Both mantle composition and lithosphere thickness vary spatially and temporally, with systematic differences between mid‐ocean ridges, ocean islands and large igneous provinces. For crystallization temperatures to provide robust evidence of mantle temperature variability, the controls of lithosphere thickness and mantle lithology on crystallization temperature must be isolated. We develop a multi‐lithology melting model for predicting crystallization temperatures of magmas in both intra‐plate volcanic provinces and mid‐ocean ridges. We find that the high crystallization temperatures seen at mantle plume localities do require high mantle temperatures. In the absence of further constraints on mantle lithology or melt productivity, we cannot robustly infer variable plume temperatures between ocean‐islands and large igneous provinces from crystallization temperatures alone; for example, the extremely high crystallization temperatures obtained for the Tortugal Phanerozoic komatiite could derive from mantle of comparable temperature to modern‐day Hawaii. This work demonstrates the limit of petrological thermometers when other geodynamic parameters are poorly known.</jats:p>