Swift: An Autoregressive Consistency Model for Efficient Weather Forecasting

Summary

We introduce Swift, an autoregressive consistency model for global weather modeling with 39× faster inference than state-of-the-art diffusion baselines. For the first time, this efficiency makes it feasible to multi-step finetune a probability flow generative model on domain-specific objectives, such as the continuous ranked probability score (CRPS).

Abstract

Diffusion models offer a physically grounded framework for probabilistic weather forecasting, but their typical reliance on slow, iterative solvers during inference makes them impractical for subseasonal-to-seasonal (S2S) applications where long lead-times and domain-driven calibration are essential. To address this, we introduce Swift, a single-step consistency model that, for the first time, enables autoregressive finetuning of a probability flow model with a continuous ranked probability score (CRPS) objective. This eliminates the need for multi-model ensembling or parameter perturbations. Results show that Swift produces skillful 6-hourly forecasts that remain stable for up to 75 days, running 39× faster than state-of-the-art diffusion baselines while achieving forecast skill competitive with the numerical-based, operational IFS ENS. This marks a step toward efficient and reliable ensemble forecasting from medium-range to seasonal-scales.