While it is accepted that both the hot strength of the matrix and the amount of undissolved carbides are important for the cutting performance of high speed steels, the relative weights of their contributions are unknown. In this work, they are separately identified and a model. is presented that provides a quantitative prediction of tool life (solely in uninterrupted cutting) on the basis of micro-structural and compositional data over a wide range of alloy compositions and cutting speeds. The model seems to describe the individual contributions to tool life well enough to serve as a guide in alloy development. The model has been developed using 13 different steels, spanning the entire range of customary compositions. It is based on the following parameters: volume fractions and compositions of undissolved carbides; precipitates formed during tempering (secondary hardening) and during operation (tertiary precipitates); and, finally, residual solute in the matrix. Tool life is modeled as a linear combination of contributions from the undissolved carbides and from the precipitate population, including a contribution due to the action of Co, and with an additional term due to solute strengthening of the matrix. The weight factors are determined by multiple linear regression analysis. They reflect the relative importance of each contributing factor, and their dependence on cutting speed can be interpreted in terms of the change in operative wear mechanism with tool temperature.