Abstract We present a rest-frame UV–optical (λ = 2500–6400 Å) stacked spectrum representative of massive quiescent galaxies at 1.0 < z < 1.3 with log(M _*/M _⊙) > 10.8. The stack is constructed using VANDELS survey data, combined with new KMOS observations. We apply two independent full-spectral-fitting approaches, measuring a total metallicity [Z/H] = −0.13 ± 0.08 with Bagpipes and [Z/H] = 0.04 ± 0.14 with Alf, a fall of ∼0.2–0.3 dex compared with the local universe. We also measure an iron abundance [Fe/H] = −0.18 ± 0.08, a fall of ∼0.15 dex compared with the local universe. We measure the alpha enhancement via the magnesium abundance, obtaining [Mg/Fe] = 0.23 ± 0.12, consistent with galaxies of similar mass in the local universe, indicating no evolution in the average alpha enhancement of log(M _*/M _⊙) ∼ 11 quiescent galaxies over the last ∼8 Gyr. This suggests the very high alpha enhancements recently reported for several bright z ∼ 1–2 quiescent galaxies are due to their extreme masses, log(M _*/M _⊙) ≳ 11.5, in accordance with the well-known downsizing trend, rather than being typical of the z ≳ 1 population. The metallicity evolution we observe with redshift (falling [Z/H], [Fe/H], constant [Mg/Fe]) is consistent with recent studies. We recover a mean stellar age of 2.5 − 0.4 + 0.6 Gyr, corresponding to a formation redshift z form = 2.4 − 0.3 + 0.6 . Recent studies have obtained varying average formation redshifts for z ≳ 1 massive quiescent galaxies, and, as these studies report consistent metallicities, we identify models with different star formation histories as the most likely cause. Larger spectroscopic samples from upcoming ground-based instruments will provide precise constraints on ages and metallicities at z ≳ 1. Combining these with precise stellar mass functions for z > 2 quiescent galaxies from the James Webb Space Telescope will provide an independent test of formation redshifts derived from spectral fitting.