TimesFM-3: A zero-shot foundation model for multivariate forecasting

Google Research introduced TimesFM-3, a 330-million-parameter time series foundation model that performs multivariate forecasting in a single forward pass, pre-trained on over 1 trillion time points. The model, released on Hugging Face, significantly outperforms other forecasting models across major benchmarks and supports zero-shot multivariate scenarios without task-specific fine-tuning.

August 31, 2026 Ayush Jain and Rajat Sen, Research Scientists, Google Research We introduce TimesFM-3, a state-of-the-art time series foundation model that enables highly accurate multivariate time series forecasting in a single forward pass, significantly outperforming other forecasting models across major benchmarks. Since the debut of TimesFM https://research.google/blog/a-decoder-only-foundation-model-for-time-series-forecasting/ in 2024, we’ve seen the adoption of time-series foundation models for real-world time-series forecasting tasks across multiple domains, such as retail, finance, observability, manufacturing, healthcare and natural sciences https://icml-structured-fm-workshop.github.io/ . Up until TimesFM-2.5 https://huggingface.co/google/timesfm-2.5-200m-pytorch released in September 2025 , our models were strictly limited to univariate forecasting: forecasting using only the history of a single time series. Yet, most real-world forecasting problems are inherently multivariate: where multiple time series and auxiliary external features jointly impact the future forecast of a time series. Consider forecasting ice cream sales for a retail chain. Past sales alone rarely tell the full story. A good forecast should also draw on sales of related products e.g., ice cream cones, syrups , historical foot traffic, and known future events like weather forecasts, promotions, and holidays. Today we introduce TimesFM-3 https://huggingface.co/google/timesfm-3.0-pytorch , the next generation of our time-series foundation model that is natively pre-trained for multivariate forecasting. TimesFM-3 has 330 million parameters and is pre-trained on a real-world and synthetic time-series corpus comprising more than 1 trillion time points. Building on the efficiency and zero-shot generalization of its predecessors, TimesFM-3 adds robust support for complex multivariate scenarios in a zero-shot manner. It can jointly predict multiple coevolving time series, capturing dependencies that improve overall accuracy without requiring task-specific fine-tuning. The model natively supports: TimesFM-3 builds on the proven decoder-only transformer architecture https://research.google/blog/transformer-a-novel-neural-network-architecture-for-language-understanding/ of its predecessors. As in previous versions, we process time series efficiently by grouping contiguous data points into patches of 32 time steps. We then apply normalization per time-series similar to that of TimesFM-2.5 https://github.com/google-research/timesfm/blob/master/src/timesfm/flax/util.py L43 in order to account for time series with vastly different scales. For target and past-covariate series, a token is constructed directly from a single patch. However, for past-future covariates, TimesFM-3 employs a clever "lookahead" strategy: each token concatenates the current patch with future patches, allowing the model to peek at upcoming known signals. Once the patches are tokenized, they pass through an input residual block and enter the main transformer stack, which operates as a 2D grid: These two attention mechanisms alternate for several layers, seamlessly blending temporal patterns with cross-series relationships. Previous versions of TimesFM generated forecasts one patch at a time, introducing latency, compounding error accumulation, and computational cost. TimesFM-3 uses the strategy of Contiguous Patch Masking https://arxiv.org/abs/2505.23719 to generate the entire forecasting horizon in a single forward pass. The model appends masked placeholder tokens for the future horizon alongside the observed context. Target and past-covariate series are masked in the horizon since their future values are unknown , while past–future covariates remain visible, providing the model with known future signals like holidays or scheduled events. Through the alternating attention layers, the model fills in all masked horizon patches simultaneously, with no iterative loop required. The model predicts 9 quantiles from the 10th to the 90th percentile for each target time series at every horizon step, providing a full probabilistic view of the forecast uncertainty. Let’s revisit the ice cream sales example. Imagine you are working on next month’s promotion schedule and want to forecast the sales to anticipate. A standard univariate model the red line, below looks at the historical sales and projects a weekly pattern forward — but it has no idea about planned promotions on specific days. TimesFM-3's multivariate mode the blue line, below takes a different approach: by passing in the planned promotion schedule as a past-future covariate, the model learns the relationship between promotions and sales lift from the historical context, then applies that knowledge to future days with planned promotions. The result is a forecast that anticipates a ~20% sales bump on each promotion day. In the chart below, the amber blocks in the promotion covariates highlight which days have promotions — and the blue forecast visibly responds to each one, while the red forecast does not. Over the full month, this adds up to a more accurate forecast for projected revenue. We evaluated TimesFM-3 on three comprehensive public forecasting benchmarks: Gift-Eval https://huggingface.co/spaces/Salesforce/GIFT-Eval , FEV-Bench https://huggingface.co/spaces/autogluon/fev-bench , and Time https://huggingface.co/spaces/Real-TSF/TIME-leaderboard . On all three benchmarks, TimesFM-3 is the top-ranked model in terms of both point and probabilistic forecasting metrics among all pre-trained foundation models. The plots below show average rank across tasks for both point forecast accuracy and probabilistic forecast quality lower is better for the three benchmarks. We compare against recent foundation models including multivariate-capable models, such as Chronos-2 and the Toto 2.0 family, as well as our previous model TimesFM-2.5. Each plot includes two entries for TimesFM-3. The "univariate mode" point shows performance when the model is evaluated without any covariate or cross-series information, treating each target series independently, just like a traditional univariate model. Even in this univariate mode, TimesFM-3 already matches or outperforms other competing models. When we switch to the full multivariate mode, TimesFM-3 takes another leap, achieving the best average rank in both point and probabilistic forecasting across the board. We introduce TimesFM-3, the latest generation of our TimesFM family of zero-shot time series foundation models, that obtains state-of-the-art multivariate and univariate forecasting performance on multiple public benchmarks. TimesFM-3 is now available on GitHub https://github.com/google-research/timesfm and Hugging Face https://huggingface.co/google/timesfm-3.0-pytorch , with its BigQuery integration landing in the coming weeks. In the meantime, you can try TimesFM-2.5 immediately on your univariate tasks to familiarize yourself with the AI.FORECAST https://docs.cloud.google.com/bigquery/docs/reference/standard-sql/bigqueryml-syntax-ai-forecast command in BigQuery - no ML expertise required. This project is joint work with Yichen Zhou, Petros Mol, Abhimanyu Das and Samet Oymak.