<p>Boundary layered stagnation point flow and thermal radiation transfer across a nonlinear stretching /shrinking sheet in hybrid carbon nanotubes (HCNTs) examines the fluid dynamics and thermal transport in the regions near stagnation induced by complex surface dynamics, the effect of nonlinear stretching, and the enhanced thermal conductivity of HCNTs, and has important consequences in the current cooling technologies, nanocating procedures, and highly efficient thermal exchangers. This communication implements Lobatto collocation information-driven nonlinear autoregressive exogenous deep neural networks (LC-ARX-DNNs) optimized with the Levenberg–Marquardt method to explore heat transfer and boundary layer response at the stagnation point of HCNTs-based magnetohydrodynamic (MHD), i.e., (HCNTs-MHD), nanofluid model over a nonlinear stretching/shrinking surface. The fourth order Lobatto IIIA collocation with residual error control technique is used to produce the information by systematically varying key parameters, including the hydrodynamic ratio parameter, shear-thinning index, electromagnetic conductivity-density parameter, nonlinear MHD parameter, thermal enhancement parameter, Prandtl number, and non-dimensional stretching/shrinking sheet to comprehensively evaluate the dynamical behavior of the nanofluidic system. The collected numerical information is employed as targets for LC-ARX-DNNs learning, based on the training, validation, and test processes, to determine the optimum solution of the HCNTs-MHD model for various physical scenarios. The efficacy, precision, and worth of the proposed technique are substantiated through learning curves over MSE ranging from 10^–09 to 10^–10, regression metric with magnitude close to unity, error histograms with an average value of the centralized bin within 10^–06–10^–07, and absolute error about 10^–04–10^–08 for the solution profiles of the HCNTs-MHD model.</p>