mlsanityData to 7 Oct 2026

Toto 2.0: Time Series Forecasting Enters the Scaling Era

arXiv 2605.20119 ↗First posted 19 May 2026cs.LG cs.AIcited by 21

Not on the map. Why?

Abstract (arXiv)

We show that time series foundation models scale: a single training recipe produces reliable forecast-quality improvements from 4M to 2.5B parameters. We release Toto 2.0, a family of five open-weights forecasting models trained under this recipe. The Toto 2.0 family sets a new state of the art on three forecasting benchmarks: BOOM, our observability benchmark; GIFT-Eval, the standard general-purpose benchmark; and the recent contamination-resistant TIME benchmark. This report describes our experimental results and details the design decisions behind Toto 2.0: its architecture and training recipe, training data, and the u-muP hyperparameter transfer pipeline. All five base checkpoints are released under Apache 2.0.

Key facts. Cited by 21 papers in the map (data through 7 Oct 2026). Builds on 32 of its 47 references that are in the map. First posted on arXiv 19 May 2026.

Builds on

32 of 47 references.

References · 32

  1. 2018BERT: Pre-training of Deep Bidirectional Transformers for Language Understandingcited by 40,204
  2. 2020Scaling Laws for Neural Language Modelscited by 6,411
  3. 2016XGBoost: A Scalable Tree Boosting Systemcited by 5,120
  4. 2019Optuna: A Next-generation Hyperparameter Optimization Frameworkcited by 2,455
  5. 2022A Time Series is Worth 64 Words: Long-term Forecasting with Transformerscited by 1,981
  6. 2024Chronos: Learning the Language of Time Seriescited by 712
  7. 2024MiniCPM: Unveiling the Potential of Small Language Models with Scalable Training Strategiescited by 694
  8. 2023A decoder-only foundation model for time-series forecastingcited by 650
  9. 2025Muon is Scalable for LLM Trainingcited by 381
  10. 2021Transformers Can Do Bayesian Inferencecited by 338
  11. 2024xLSTM: Extended Long Short-Term Memorycited by 320
  12. 2020Query-Key Normalization for Transformerscited by 266
20 more
  1. 2024GIFT-Eval: A Benchmark For General Time Series Forecasting Model Evaluationcited by 188
  2. 2025Chronos-2: From Univariate to Universal Forecastingcited by 158
  3. 2023Tree-Structured Parzen Estimator: Understanding Its Algorithm Components and Their Roles for Better Empirical Performancecited by 143
  4. 2017Dissecting Adam: The Sign, Magnitude and Variance of Stochastic Gradientscited by 136
  5. 2025TiRex: Zero-Shot Forecasting Across Long and Short Horizons with Enhanced In-Context Learningcited by 107
  6. 2025The Polar Express: Optimal Matrix Sign Methods and Their Application to the Muon Algorithmcited by 104
  7. 2025NorMuon: Making Muon more efficient and scalablecited by 74
  8. 2024Time-MMD: Multi-Domain Multimodal Dataset for Time Series Analysiscited by 65
  9. 2025Moirai 2.0: When Less Is More for Time Series Forecastingcited by 59
  10. 2025This Time is Different: An Observability Perspective on Time Series Foundation Modelscited by 57
  11. 2024Toto: Time Series Optimized Transformer for Observabilitycited by 44
  12. 2024u-μP: The Unit-Scaled Maximal Update Parametrizationcited by 29
  13. 2026It's TIME: Towards the Next Generation of Time Series Forecasting Benchmarkscited by 25
  14. 2023Unit Scaling: Out-of-the-Box Low-Precision Trainingcited by 17
  15. 2025Time-IMM: A Dataset and Benchmark for Irregular Multimodal Multivariate Time Seriescited by 16
  16. 2025TempoPFN: Synthetic Pre-training of Linear RNNs for Zero-shot Time Series Forecastingcited by 16
  17. 2025Cautious Weight Decaycited by 12
  18. 2025Decoder-Hybrid-Decoder Architecture for Efficient Reasoning with Long Generationcited by 10
  19. 2025Fidel-TS: A High-Fidelity Multimodal Benchmark for Time Series Forecastingcited by 5
  20. 2026ARFBench: Benchmarking Time Series Question Answering Ability for Software Incident Responsecited by 5

Built on by

Cited by year (2026 to 7 Oct)

Citing papers per year: 2026: 21.

21 papers.

Sort:

By area:

  1. 2026Falcon-X: A Time Series Foundation Model for Heterogeneous Multivariate Modelingcited by 7
  2. 2026Into the ORBIT for Time Series: Training Regimes for Foundation Modelscited by 3
  3. 2026EXAONE Finance 1.0: An Attention-free Time Series Foundation Model for Financial Time Seriescited by 2
  4. 2026TinyCast: Probabilistic Zero-Shot Forecasting with Computed Periodicitycited by 1
  5. 2026Tabby: An Open Pretraining Recipe for Time Series Foundation Modelscited by 1
  6. 2026Sharp Stability Threshold and Certification for Designing Stable Residual Architecturescited by 0
  7. 2026LLM-based Agents for Forecasting and Prediction: Methods, Training, Evaluation, and Applicationscited by 0
  8. 2026Network-Aware Forecasting on Wireless Access Pointscited by 0
  9. 2026Distillation of Synthetic Data for Time Series Foundation Modelscited by 0
  10. 2026Beyond Numerical Time Series: A Unified Benchmark for Multimodal Forecasting with Heterogeneous Contextcited by 0
  11. 2026t0: A Time-Series Foundation Model for Forecasting with Contextcited by 0
  12. 2026When Tomorrow Becomes Today: Self-Evolving Policies for Agentic Time-Series Forecastingcited by 0
9 more
  1. 2026Aurora-X: Built for Extreme Time Series Forecastingcited by 0
  2. 2026Self-Evolving Time-Series Forecasting Agents with Episodic Memory and Online Policy Learningcited by 0
  3. 2026Instance-Adaptive Prompts as Context for Time-Series Foundation Modelscited by 0
  4. 2026JudgeCast: Time Series Forecasting with Experience-Informed Covariate Judgementscited by 0
  5. 2026The Row Normalization Puzzle in Muoncited by 0
  6. 2026When, Not How Much: Evaluating Time-Series Foundation Models on Sparse Eventscited by 0
  7. 2026OpenTSLM TeeMoE: A Unified Time-Series Language Model for Forecasting, Contextual Prediction, and Reasoningcited by 0
  8. 2026Timer-M1: A Multivariate Time Series Foundation Model via Learning Primitivescited by 0
  9. 2026AdaCast: Conditional Parameter Generation for Adaptive Time Series Forecastingcited by 0