Timestral
AssessPlatforms
A family of time-series foundation models for zero-shot forecasting and domain adaptation.
Why it's here
Placed in Assess: 1 article(s) of evidence from 1 source(s), led by model releases, with 1 in the last 30 days. Confidence 24%. Low accumulated evidence, so it defaults conservatively pending more signal.
Evidence (1)
- 7Hacker News·10/19/2026model_releaseMistral launches Timestral time-series foundation models
Mistral introduced Timestral, a family of time-series foundation models designed for probabilistic, multivariate forecasting with support for past and future covariates. The largest 209M-parameter model claims top results on GIFT-Eval, TIME, and fev-bench while running efficiently on CPUs and supporting fine-tuning through Mistral Forge.