Carbon removals associated with incremental gains in soil organic carbon (SOC) at scale have enormous potential to mitigate global warming, yet confusion over contexts that elicit SOC accrual abound. Here, we examine how bespoke interventions (through irrigation, fertiliser, crop type and rotations), antecedent SOC levels and soil type impact on long-term SOC accrual and greenhouse gas (GHG) emissions. Using a whole farm systems modelling approach informed using participatory research, we discovered an inverse relationship between antecedent SOC stocks and SOC gains realised following intervention, with greater initial SOC levels resulting in lower ex poste change in SOC. We found that SOC accrual was greatest for clays and least for sands, although changes in SOC in sandy loam soils were also low. Diversified whole farm adaptations – implemented through inclusion of grain legumes within wheat/canola crop rotations – were more conducive to improvement in SOC stocks, followed by Intensified systems (implemented through greater rates of irrigation, farm areas under irrigation, nitrogen fertiliser and inclusion of rice and maize in crop rotations). Adaptations that Simplified farm systems by reducing irrigation and fertiliser use resulted in the lowest SOC accrual. In most cases, long-term SOC stocks fell when SOC at the outset was greater than 4–5%, regardless of intervention made, soil or crop type, crop rotation, production system or climate. We contend that (1) management interventions primarily impacted SOC in the soil surface (0–30 cm) and had de minimus impact on deep SOC stocks (30–100 cm), (2) crop rotations including wheat, canola and faba beans were more conducive to improvement in SOC stocks, (3) scenarios with high status quo SOC had little impact on crop productivity, and not necessarily the lowest GHG emissions intensity, (4) productivity and GHG emissions intensity were largely a function of the quantum of nitrogenous fertiliser added, rather than SOC stocks, and (5) aspirations for improving SOC are likely to be futile if antecedent SOC stocks are already high (4–5 %). We conclude that potential for improving SOC stocks exists in contexts where antecedent stocks are low (<1%), which may include regions with land degradation, chronic erosion and/ or other constraints to vegetative ground cover that could be sustainably and consistently alleviated.