Garnet-group minerals, with their wide range of compositions, play a significant role in Earth's crust and upper mantle, participating in various petrological and geochemical processes. Oxygen isotope fractionation factors between garnets and other minerals hold crucial implications in these contexts. In this work, vibration frequencies were measured via Raman spectra on five garnet mineral samples in the pyrope-almandine-spessartine ternary system and four synthetic pyrope-grossular solid solutions at temperatures up to 1000 °C and pressures up to 17 GPa. Isobaric (γ_iP) and isothermal (γ_iT) mode Grüneisen, as well as anharmonic (a_i) parameters, were determined for the observed modes. Our study shows that the vibration bands tend to shift to lower frequencies with increasing temperature and to higher frequencies with increasing pressure. Moreover, the anharmonic correction has been found to contribute positively to equilibrium oxygen isotope fractionation β factors in garnets, typically within 0.5 ‰ above 1000 K. The 10³∙lnβ(T) factors calculated from vibration spectra (considering Г points only) align closely with those from theoretical calculations (including all frequencies in the Brillouin zone). Contributions from phonon dispersion (including none-Г points) are typically smaller than the uncertainties propagated from frequency measurements. Additionally, the pressure effect on the oxygen isotope fractionation in garnet, evaluated from the isothermal Grüneisen parameters, is found to be insignificant under crustal and upper mantle conditions, consistent with thermodynamic expectations. Using published oxygen isotope fractionation β factors for quartz and calcite, the equilibrium oxygen isotope fractionation factors (10³∙lnα) are calculated between garnets of diverse compositions and any of these phases as functions of temperature, pressure and fraction of garnet endmembers. Considering composition effects helps to reduce the discrepancies among experimental and empirical calibrations of oxygen isotope fractionation factors between garnet and quartz. As an example application, oxygen isotope fractionation factors 10³∙lnα(P, T) between garnets and forsterite were calculated under upper mantle conditions. Our findings suggest that these fractionation factors are predominantly dependent on temperature and garnet compositions, with minor influence from pressure. Compared with our calculated fractionation factors between garnet and olivine, the oxygen isotope exchange equilibrium may have been reached in some in kimberlite and mantle xenolith samples between garnet and olivine. However, in other samples, the fractionation factors cannot be explained solely by the composition effect, likely due to metasomatism. Our results underscore the importance of considering garnet compositions when interpreting their oxygen isotope compositions.