Soils are indispensable for the provision of several functions. Agricultural intensification and its focus on increasing primary productivity (PP) poses a threat to soil quality, due to increases in nutrient loads, greenhouse gas emissions and declining biodiversity. The EU Horizon 2020 Landmark project has developed multi-criteria decision models to assess five soil functions: PP, nutrient cycling (NC), soil biodiversity and habitat provision (B-HP), climate mitigation and water regulation, simultaneously in agricultural fields. Using these algorithms, we evaluated the supply of PP, NC and B-HP of 31 grasslands and 21 croplands as low, medium or high. The multi-criteria decision models showed that 38% of the farms had a medium to high supply of all three soil functions, whereas only one cropland had a high supply for all three. Forty-eight per cent of the farms were characterized by a high supply of PP and NC. We observed a clear trade-off between these two functions and B-HP. Multivariate statistical analyses indicated that higher organic inputs combined with a lower mineral fertilization concur with higher biodiversity scores while maintaining a medium delivery of PP and NC. Additionally, we compared the outputs of the model predictions to independent variables that served as proxies for the soil functions and found: (a) croplands (but not grasslands) with high PP had a higher standardized yield than those with medium PP; (b) grasslands (but not croplands) with high NC had a significantly lower fungal to bacterial biomass ratio, suggesting faster decomposition channels; and (c) a positive though non-significant trend between B-HP score and rank according to soil invertebrate biodiversity. These comparisons suggest a successful upscaling of the models from field to farm level. Our study highlights the need for systematic collection of management-related data for the assessment of soil functions. Multifunctionality can be achieved in agricultural soils; however, without specifically managing for it, biodiversity might come at a loss.