While significant advances have come about for turbomachinery off-design performance characterisation using computational fluid dynamics, the need for quick performance estimates at challenging off-design conditions still requires the use of lower-order models, such as mean-line analyses and through-flow tools. These inviscid tools require blade performance correlations formulated in terms of loss and turning angle as a function of blade geometric and aerodynamic parameters. Traditionally, such correlations have relied on empirical data from blade cascade tests at nominal incidence conditions. This limitation on the applicability of the blade correlations has caused performance modelling of the sub-idle regime to be off-limits to this type of correlation-based approaches. This paper addresses the development of blade loss and deviation models applicable to the sub-idle regime using a parametric numerical approach. 2D CFD results are used to generate a model that is then applied to mean-line and through-flow analyses aimed at predicting the sub-idle map of an axial flow compressor. The model proves to be a valuable tool for quick sub-idle performance estimates and allows existing correlation-based performance prediction methods to be extended into the sub-idle regime.