Mistral Releases Timestral, Four Open Time-Series Forecasting Models Up to 209M Parameters

Mistral AI released Timestral, its first family of time-series foundation models, on October 15. The four models (5M, 28M, 123M and 209M parameters) produce probabilistic, multivariate zero-shot forecasts, with a 32,768-step context and 21 output quantiles.

Mistral says the 209M model ranks first among pretrained zero-shot models on all five reported metrics across GIFT-Eval, TIME and fev-bench, using about two-thirds of the parameters of the previous best. The 28M model reportedly beats Chronos-2 on GIFT-Eval and TIME despite being roughly four times smaller.

The models run on CPU: a one-day-ahead forecast takes about 10 ms with the 5M model and about 140 ms with the 209M model on an 8-thread server. They ship under a modified MIT license, with code on GitHub and weights on Hugging Face, and can be deployed on-premises, in a VPC or air-gapped.