← Tuesday, October 20, 2026Models & trainingRelease

Mistral’s Timestral forecasts time series with models up to 209M parameters

mistral.ai 7 min read Editor score 7/10

Mistral says its 209M-parameter model ranks first on GIFT-Eval, TIME and fev-bench, using synthetic pretraining, covariate-aware attention and long-context learning. The reported results make this release worth assessing for forecasting workloads and CPU deployment.

What

Timestral is Mistral’s family of time-series foundation models, with four sizes from 5M to 209M parameters.

How

It uses synthetic pre-training, covariate-aware attention and long-context learning to produce probabilistic multivariate forecasts, including zero-shot forecasts.

Why it matters

The family targets forecasting across domains and offers smaller, faster, cheaper models alongside larger models with better forecast performance.

Read on mistral.ai → Hacker News · 106 comments

Headline, summary, brief and diagram are written by a model from the source's own text; the link, byline and numbers come from the source. Read the original before relying on a claim.