Abstract Most known porphyry Cu ± Au deposits are associated with moderately oxidized and sulfur-rich, calc-alkaline to mildly alkalic arc-related magmas in the Phanerozoic. In contrast, sodium-enriched tonalite–trondhjemite–granodiorite–diorite (TTG) magmas predominant in the Archean are hypothesized to be unoxidized and sulfur-poor, which together preclude porphyry Cu deposit formation. Here, we test this hypothesis by interrogating the causative magmas for the ~2·7 Ga TTG-related Côté Gold, St-Jude, and Clifford porphyry-type Cu ± Au deposit settings in the Neoarchean southern Abitibi subprovince. New and previously published geochronological results constrain the age of emplacement of the causative magmas at ~2·74 Ga, ~2·70 Ga, and ~ 2·69 Ga, respectively. The dioritic and trondhjemitic magmas associated with Côté Gold and St-Jude evolved along a plagioclase-dominated fractionation trend, in contrast to amphibole-dominated fractionation for tonalitic magma at Clifford. Analyses of zircon grains from the Côté Gold, St-Jude, and Clifford igneous rocks yielded εHf(t) ± SD values of 4·5 ± 0·3, 4·2 ± 0·6, and 4·3 ± 0·4, and δ18O ± SD values of 5·40 ± 0·11 ‰, 3·91 ± 0·13 ‰, and 4·83 ± 0·12 ‰, respectively. These isotopic signatures indicate that, although these magmas are mantle-sourced with minimal crustal contamination, for the St-Jude and Clifford settings the magmas or their sources may have undergone variable alteration by heated seawater or meteoric fluids. Primary barometric minerals (i.e. zircon, amphibole, apatite, and magnetite–ilmenite) that survived variable alteration and metamorphism (up to greenschist facies) were used for estimating fO2 of the causative magmas. Estimation of magmatic fO2 values, reported relative to the fayalite–magnetite–quartz buffer as ΔFMQ, using zircon geochemistry indicates that the fO2 values of the St-Jude, Côté Gold, and Clifford magmas increase from ΔFMQ –0·3 ± 0·6 to ΔFMQ +0·8 ± 0·4 and to ΔFMQ +1·2 ± 0·4, respectively. In contrast, amphibole chemistry yielded systematically higher fO2 values of ΔFMQ +1·6 ± 0·3 and ΔFMQ +2·6 ± 0·1 for Côté Gold and Clifford, respectively, which are consistent with previous studies that indicate that amphibole may overestimate the fO2 of intrusive rocks by up to 1 log unit. Micro X-ray absorption near edge structure (μ-XANES) spectrometric determination of sulfur (i.e. S6+/ΣS) in primary apatite yielded ≥ΔFMQ −0·3 and ΔFMQ +1·4–1·8 for St-Jude and Clifford, respectively. The magnetite–ilmenite mineral pairs from the Clifford tonalite yielded ΔFMQ +3·3 ± 1·3 at equilibrium temperatures of 634 ± 21 °C, recording the redox state of the late stage of magma crystallization. Electron probe microanalyses revealed that apatite grains from Clifford are enriched in S (up to 0·1 wt%) relative to those of Côté Gold and St-Jude (below the detection limit), which is attributed to either relatively oxidized or sulfur-rich features of the Clifford tonalite. We interpret these results to indicate that the deposits at Côté Gold and Clifford formed from mildly (~ΔFMQ +0·8 ± 0·4) to moderately (~ΔFMQ +1·5) oxidized magmas where voluminous early sulfide saturation was probably limited, whereas the St-Jude deposit represents a rare case whereby the ingress of externally derived hydrothermal fluids facilitated metal fertility in a relatively reduced magma chamber (~ΔFMQ +0). Furthermore, we conclude that variable modes of formation for these deposits and, in addition, the apparent rarity of porphyry-type Cu–Au deposits in the Archean may be attributed to either local restriction of favorable metallogenic conditions, and/or preservation, or an exploration bias.