Terahertz frequency quantum cascade lasers offer a potentially vast number of new applications. To better understand and apply these lasers, a device-specific modeling method was developed that realistically predicts optical output power under changing current drive and chip temperature. Model parameters are deduced from the self-consistent solution of a full set of rate equations, obtained from energy-balance Schro ̈dinger-Poisson scattering transport calculations. The model is thus derived from first principles, based on the device structure, and is therefore not a generic or phenomenological model that merely imitates expected device behavior. By fitting polynomials to data arrays representing the rate equation parameters, we are able to significantly condense the model, improving memory usage and computational efficiency.