The M1-multigroup model intricately captures the interaction between light and matter, incorporating the spectral behaviour of photons. However, a critical aspect of this model lies in determining the Eddington factor, which is used in the closure relation linking radiative pressure to radiative energy. Despite lacking an analytical expression, our investigation unveils that the Eddington factor depends solely on three parameters: the radiative temperature, the reduced flux, and the frequency bounds’ ratio of a considered group. To address this challenge, we have devised a novel approach leveraging neural networks and polynomials, enabling rapid and accurate estimation of this quantity. Our method showcases significant advantages over existing techniques. It demonstrates speeds up to 3,000 times faster than the most precise method utilising a line search algorithm and achieves precision levels up to 1,000 times higher than those relying on the M1-grey model’s Eddington factor expression. Moreover, unlike interpolation-based methods, our approach eliminates the need for prior knowledge of radiative quantities. Consequently, our method emerges as one of the most efficient means to precisely compute the Eddington factor in the M1-multigroup model, offering a potent tool for advancing radiative hydrodynamics simulations.