> ## Content Index
> Fetch the complete content index at: https://www.ai-360.online/llms.txt
> Use this file to discover other available public pages before exploring further.

# Mistral AI releases Timestral time-series models
- URL: https://www.ai-360.online/mistral-ai-introducing-timestral/
- Published: 2026-10-15T12:00:00.000Z
- Updated: 2026-10-31T20:47:40.000Z
- Description: Mistral AI's Timestral time-series models, from 5M to 209M parameters, run on CPUs and top GIFT-Eval, TIME and fev-bench forecasting benchmarks.
- Author: Nicholas Lipinski
- Tags: 1. Enterprise AI, News, #company-mistral-ai

[Mistral AI](https://mistral.ai/?ref=ai-360.online) has released Timestral, its first family of time-series foundation models, for probabilistic, multivariate forecasting, [the company said on 15 July 2026](https://mistral.ai/news/timestral?ref=ai-360.online).

The models come in four sizes, from 5 million to 209 million parameters, and run efficiently on CPUs. [Mistral](https://www.ai-360.online/mistral-ai/) said the family sits on the Pareto front of accuracy versus model size, and the largest ranks first on GIFT-Eval, TIME and fev-bench, beating the previous state of the art with two-thirds of the parameters.

Timestral supports past-only and known-future covariates and zero-shot forecasting. It combines synthetic pre-training, covariate-aware attention and long-context learning.

Organisations can fine-tune it for their own domains through Mistral Forge, the training pipeline Mistral uses for its language models.

The announcement was published on 15 July 2026\. Mistral did not give pricing for Forge fine-tuning.

---