TimesFM-3
TimesFM-3 is Google Research's 330-million-parameter time-series foundation model for zero-shot univariate and multivariate forecasting.
Provider
Model family
Google TimesFM
Time-series foundation model
Cost tier
Timesfm 3
Status
Current
Release Aug 31, 2026
Why teams choose it
Google Research says TimesFM-3 is natively pretrained for multivariate forecasting
unlike TimesFM-2.5 and earlier checkpoints which were univariate.
The Hugging Face checkpoint ID is google/timesfm-3.0-pytorch
The model predicts nine quantiles per target step and fills the horizon in one forward pass.
Source code remains Apache 2.0. TimesFM-3 pretrained weights use timesfm-non-commercial-license-v1.0
TimesFM-2.5 weights remain Apache 2.0 and are the current BigQuery AI.FORECAST path until 3.0 lands.
Tradeoffs to know
- Commercial or production use of the default TimesFM-3 pretrained weights is not permitted under the published non-commercial license.
- Benchmark ranks versus Chronos-2, Toto 2.0, and TimesFM-2.5 are vendor-reported on GIFT-Eval, FEV-Bench, and TIME and should be reproduced on your series.
When not to use this
- BigQuery TimesFM-3 integration was announced as coming in the following weeks; do not assume 3.0 is live in AI.FORECAST yet.
Technical specs
- Inputs
- time-series
- Outputs
- time-series
- Capabilities
- univariate forecasting, multivariate forecasting, zero-shot forecasting, past covariates, past-future covariates, quantile forecasts, single-pass decoding
- License
- TimesFM Non-Commercial License v1.0 (pretrained weights); Apache 2.0 (code)
- Source
- https://research.google/blog/timesfm-3-a-zero-shot-foundation-model-for-multivariate-forecasting/
- Fev Bench
- top pretrained foundation model
- Gift Eval
- top pretrained foundation model
- Parameters
- 330M
- Pretrain Time Points
- 1T+
Benchmarks
- Time Leaderboard
- top pretrained foundation model
Related models
Gemma 2 27B
Gemma 2 27B is Google’s open-weights Gemma family checkpoint balancing quality and deployability for research and pro…
Gemini 3.1 Pro Preview
Gemini 3.1 Pro Preview is Google's preview Gemini 3-series model for complex multimodal reasoning, software engineeri…
Gemini 2.0 Flash
Gemini 2.0 Flash is Google’s efficiency-oriented multimodal model generation aimed at fast agentic and interactive ex…
TimesFM-3 FAQ
What is TimesFM-3?
TimesFM-3 is Google Research's 330-million-parameter time-series foundation model for zero-shot univariate and multivariate forecasting. The August 31, 2026 announcement documents native multivariate forecasting in a single forward pass, past and past-future covariates, a pretraining corpus of more than 1 trillion time points, and availability on GitHub and Hugging Face, with BigQuery integration planned in the coming weeks.
When does TimesFM-3 fit best?
Retail and demand forecasting with related series
What should teams watch out for with TimesFM-3?
Commercial or production use of the default TimesFM-3 pretrained weights is not permitted under the published non-commercial license.
Explore next
Models, tools, and comparisons that connect to this reference.