Timestral
A family of time-series foundation models released by Mistral, based on a zero-shot forecasting paradigm. It supports CPU inference, can output multi-quantile probabilistic forecasts, and accepts covariate inputs.
Timeline (last 90 days)
Mistral released Timestral, a family of four time series foundation models (5M, 28M, 123M, 209M) sharing the same architecture and training recipe.
The Timestral models are zero-shot, so users can make predictions without training on their own data.
The largest Timestral version (209M parameters) ranks first on three major public benchmarks.
Timestral does not require a GPU for inference.
The Timestral weights are released under a modified MIT license, with the restriction that a company is not authorized to use them if its global monthly revenue last month exceeded $20 million.
The Timestral inference code is under the Apache 2.0 license, while the weights use a modified MIT license.
Mistral says Timestral's results come from a single checkpoint, with no ensemble, and have been submitted to the GIFT-Eval and TIME leaderboards, and that most of the training data is synthetic.
The Timestral documentation lists Timestral 1.0 as a public preview and does not state API pricing.
Installing Timestral requires installing from GitHub, requires Python 3.10+ and torch 2.4+, and the package is not on PyPI.
Mistral positions Timestral for local, VPC, and air-gapped deployments.
1 more timeline events
All Facts (11)
Timestral supports inputting multiple series and covariates (known in the past or known in the future) at the same time, with missing values passed in as NaN.
50%UnverifiedThe Timestral models are zero-shot, so users can make predictions without training on their own data.
50%UnverifiedThe largest Timestral version (209M parameters) ranks first on three major public benchmarks.
50%UnverifiedTimestral does not require a GPU for inference.
50%UnverifiedThe Timestral weights are released under a modified MIT license, with the restriction that a company is not authorized to use them if its global monthly revenue last month exceeded $20 million.
50%UnverifiedThe Timestral inference code is under the Apache 2.0 license, while the weights use a modified MIT license.
50%UnverifiedMistral says Timestral's results come from a single checkpoint, with no ensemble, and have been submitted to the GIFT-Eval and TIME leaderboards, and that most of the training data is synthetic.
50%UnverifiedThe Timestral documentation lists Timestral 1.0 as a public preview and does not state API pricing.
50%UnverifiedInstalling Timestral requires installing from GitHub, requires Python 3.10+ and torch 2.4+, and the package is not on PyPI.
50%UnverifiedMistral positions Timestral for local, VPC, and air-gapped deployments.
50%UnverifiedMistral released Timestral, a family of four time series foundation models (5M, 28M, 123M, 209M) sharing the same architecture and training recipe.
50%