With low-temperature combustion engine research reaching an applicable level, physics-based control-oriented models regain attention. For reactivity controlled combustion concepts, chemical kinetics-based multizone models have been proven to reproduce the governing physics for performance-oriented simulations. They offer accuracy levels similar to high-fidelity computational fluid dynamics (CFD) models but with a fraction of their computational effort. Nevertheless, state-of-the-art reactivity controlled compression ignition (RCCI) simulations with multizone model toolchains still face challenges related to predictivity and calculation speed. This study introduces a new multizone modelling framework that addresses these challenges. It includes a C++ code, deeply integrated with open-source, thermo-kinetic libraries, and coupled to an industry standard 1-D modelling framework. Incorporating a predictive turbulence mixing model, it aims to eliminate dependence on CFD-based initialisation, while applying a novel zonal configuration to achieve sensitivity to the combustion chamber´s geometrical features. Basic sensitivity simulations performed for zonal resolution and chemical kinetic mechanisms prove the approach is fit for purpose. Aiming for optimal trade-off between accuracy and simulation speed, the 12-zone model has a simulation time below three minutes per closed cycle. These achievements are validated against a medium-speed, large-bore, single-cylinder research engine, running in a dual-fuel mode with natural gas and light fuel oil. Using basic submodels, the framework reproduces measured in-cylinder pressure trace within an RMS error of 0.85 bar, and combustion performance indicators within a 5% error margin target. Ultimately, this is the first time the multi-zone kinetic framework has been proven suitable to reproduce RCCI combustion on a state-of-the-art marine engine geometry.