{"id":"arxiv:2608.06241","kind":"preprint","title":"Timestep-Conditioned Transformers for Global Weather Forecasting","authors":["Sam Levang","Fran Bartolic","Ty Dickinson","Chase Dwelle","Paulius Rauba","Viktor Cikojevic"],"author_count":6,"year":2026,"published":"2026-08-06","venue":null,"cited_by":0,"doi":null,"arxiv_id":"2608.06241","nct_id":null,"source_url":"https://arxiv.org/abs/2608.06241","archive_url":"https://space-kitty.com/papers/arxiv:2608.06241","updated":"2026-08-06","pmid":null,"pmcid":null,"openalex_id":"W7201834324","pdf_url":"https://arxiv.org/pdf/2608.06241","abstract":"Existing machine-learning weather forecasting models rely on predetermined and fixed autoregressive timesteps. The choice of model timestep involves a fundamental trade-off: shorter timesteps (e.g. 1 to 6 hours) finely resolve atmospheric dynamics within the diurnal cycle but increase error accumulation for a given forecast horizon, while longer timesteps (e.g. 24 hours) reduce error accumulation but limit the usability of short-range forecasts where sub-daily predictability is high. In this work, we present GEM-3, a probabilistic global weather model that addresses this trade-off through explicit multi-timestep inference. With a single set of trained weights, the model timestep can be configured at inference time to balance predictability and usability across a broad forecast horizon. Additionally, we find that mixed-timestep training consistently improves rollout stability relative to timestep-specialist models. Under the hood, GEM-3 is a lightweight neighborhood-attention transformer with ~134M parameters on an equirectangular grid with a number of architectural advancements beyond its predecessor GEM-2. The result is a practical forecasting system that couples near-SOTA medium-range probabilistic skill, stable extended-range rollouts, efficient training and inference, and decision-relevant diagnostics.","abstract_license":"CC0-1.0 (arXiv descriptive metadata)","license":null,"is_open_access":false,"retracted":false,"peer_reviewed":false,"categories":["cs.LG","cs.OS"],"details":null,"provenance":"arxiv","attribution":"Abstract: arXiv descriptive metadata (CC0). Thank you to arXiv for use of its open access interoperability. This service was not reviewed or approved by, nor does it necessarily express or reflect the policies or opinions of, arXiv.","topics":[{"slug":"ai-weather-climate","name":"AI Weather & Climate Modeling","domain_slug":"energy"}],"cited_in_dossiers":[]}