mlsanityData to 14 Oct 2026

t0: A Time-Series Foundation Model for Forecasting with Context

arXiv 2609.24559 ↗First posted 21 Sep 2026cs.LG cs.AIcited by 2

Not on the map. Why?

Abstract (arXiv)

We present t0, a family of open-weights foundation models for forecasting with multivariate context. We release its first two members: textttt0-alpha and textttt0-beta, respectively 102M and 256M parameters. Both condition their forecasts on target history, past covariates, and known-future covariates, without task-specific retraining. Their transformer layers alternate attention along time and across variates. They produce probabilistic forecasts through quantile predictions. Pretraining combines curated public data with synthetic generator families constructed to contain covariate-to-target dependencies. On GIFT-Eval, textttt0-alpha reaches an aggregate CRPS of 0.4941, and textttt0-beta a CRPS of 0.4738 and a MASE of 0.6865, third on both and within 4.0% of the best zero-shot TSFM. On fev-bench they score 42.2 and 46.7 in skill, the latter third again and 2.0 points behind the leader. We analyze textttt0-alpha in depth. Known-future covariates raise its skill by 6.3 percentage points across 30 tasks. The report also examines its calibration, its rollout strategy on long horizons, and its robustness to missing data. On the Victoria electricity-demand benchmark, textttt0-beta is among the most accurate models with a context of nearly a year. In an independent Macrocosm evaluation of hourly ERCOT prices over 29 months, both cut the MAE of the lagged-price baseline by 38%.

Key facts. Cited by 2 papers in the map (data through 14 Oct 2026). Builds on 43 of its 74 references that are in the map. First posted on arXiv 21 Sep 2026.

Builds on

43 of 74 references.

References · 43

  1. 2019Exploring the Limits of Transfer Learning with a Unified Text-to-Text Transformercited by 14,599
  2. 2016XGBoost: A Scalable Tree Boosting Systemcited by 5,166
  3. 2021RoFormer: Enhanced Transformer with Rotary Position Embeddingcited by 4,703
  4. 2022A Time Series is Worth 64 Words: Long-term Forecasting with Transformerscited by 2,039
  5. 2020GLU Variants Improve Transformercited by 1,804
  6. 2019Root Mean Square Layer Normalizationcited by 1,480
  7. 2017DeepAR: Probabilistic Forecasting with Autoregressive Recurrent Networkscited by 1,013
  8. 2019N-BEATS: Neural basis expansion analysis for interpretable time series forecastingcited by 753
  9. 2023Scaling Vision Transformers to 22 Billion Parameterscited by 694
  10. 2023A decoder-only foundation model for time-series forecastingcited by 692
  11. 2024Unified Training of Universal Time Series Forecasting Transformerscited by 614
  12. 2024xLSTM: Extended Long Short-Term Memorycited by 327
31 more
  1. 2020Query-Key Normalization for Transformerscited by 280
  2. 2024GIFT-Eval: A Benchmark For General Time Series Forecasting Model Evaluationcited by 212
  3. 2017A Multi-Horizon Quantile Recurrent Forecastercited by 211
  4. 2025Chronos-2: From Univariate to Universal Forecastingcited by 194
  5. 2022A Length-Extrapolatable Transformercited by 127
  6. 2025TiRex: Zero-Shot Forecasting Across Long and Short Horizons with Enhanced In-Context Learningcited by 120
  7. 2023ForecastPFN: Synthetically-Trained Zero-Shot Forecastingcited by 104
  8. 2024ChatTS: Aligning Time Series with LLMs via Synthetic Data for Enhanced Understanding and Reasoningcited by 103
  9. 2024ChatTime: A Unified Multimodal Time Series Foundation Model Bridging Numerical and Textual Datacited by 100
  10. 2025fev-bench: A Realistic Benchmark for Time Series Forecastingcited by 64
  11. 2025From Tables to Time: Extending TabPFN-v2 to Time Series Forecastingcited by 60
  12. 2024Toto: Time Series Optimized Transformer for Observabilitycited by 48
  13. 2024Fewer Truncations Improve Language Modelingcited by 35
  14. 2024In-Context Fine-Tuning for Time-Series Foundation Modelscited by 35
  15. 2026It's TIME: Towards the Next Generation of Time Series Forecasting Benchmarkscited by 33
  16. 2021Learning Quantile Functions without Quantile Crossing for Distribution-free Time Series Forecastingcited by 32
  17. 2026Toto 2.0: Time Series Forecasting Enters the Scaling Eracited by 30
  18. 2020MQTransformer: Multi-Horizon Forecasts with Context Dependent and Feedback-Aware Attentioncited by 29
  19. 2026Timer-S1: A Billion-Scale Time Series Foundation Model with Serial Scalingcited by 28
  20. 2025ChronosX: Adapting Pretrained Time Series Models with Exogenous Variablescited by 23
  21. 2025CauKer: Classification Time Series Foundation Models Can Be Pretrained on Synthetic Datacited by 13
  22. 2026TiRex-2: Generalizing TiRex to Multivariate Data and Streamingcited by 13
  23. 2025Zero-Shot Time Series Forecasting with Covariates via In-Context Learningcited by 12
  24. 2026Falcon-X: A Time Series Foundation Model for Heterogeneous Multivariate Modelingcited by 10
  25. 2025Xihe: Scalable Zero-Shot Time Series Learner Via Hierarchical Interleaved Block Attentioncited by 8
  26. 2025CoRA: Covariate-Aware Adaptation of Time Series Foundation Modelscited by 6
  27. 2024Metadata Matters for Time Series: Informative Forecasting with Transformerscited by 5
  28. 2026TS-ICL: A Flexible Time-Indexed Foundation Model for Time Series via In-Context Learningcited by 4
  29. 2025Time to Embed: Unlocking Foundation Models for Time Series with Channel Descriptionscited by 2
  30. 2026Time-Aware Prior Fitted Networks for Zero-Shot Forecasting with Exogenous Variablescited by 2
  31. 2025Forking-Sequences: Statistically and Computationally Efficient Multi-Horizon Forecasting with Reduced Volatilitycited by 1

Built on by

Cited by year (2026 to 14 Oct)

Citing papers per year: 2026: 2.

2 papers.

  1. 2026A Unified Scaling Law for Time Series Foundation Modelscited by 0
  2. 2026On the Encoding of Calendar Features in Covariate-Informed Time Series Foundation Modelscited by 0