Researchers have enhanced the Aardvark Weather model, an end-to-end AI weather forecasting system, by incorporating probabilistic capabilities through stochastic mechanisms at each component. Unlike traditional deterministic models, this approach introduces learned, input-dependent noise at the observation encoder to capture aleatoric uncertainty and additional stochastic layers to address model uncertainty, enabling the system to provide uncertainty-aware forecasts while maintaining the efficiency of direct prediction from raw Earth observations. This development aims to improve the reliability and interpretability of AI-driven weather predictions by quantifying contributions from both observational data and model limitations.