Ministral is Mistral AI's family of small language models built to run on devices and local machines rather than in big data centres.
In plain words
Big models need powerful servers. Ministral models are small enough for laptops, phones or other edge hardware, which keeps data local and answers fast. The first pair, Ministral 3B and 8B, came out in October 2024. The Ministral 3 family followed in December 2025 in 3B, 8B and 14B sizes, with text and image understanding, under the Apache 2.0 licence.
A quick example
Imagine a translation app that must work on a plane with no internet. A small model like Ministral 3B could run on the device itself, so the text never leaves it and there is no network delay.
Where you'll meet it
- On-device assistants and offline translation.
- Cheap helper steps inside larger agent workflows.
Related
Sources
- DocsUn Ministral, des Ministraux, Mistral AI · read 29 Sept 2026
- DocsIntroducing Mistral 3, Mistral AI · read 29 Sept 2026
- DocsMinistral 3 8B, Mistral Docs · read 29 Sept 2026
Checked by WhatIsAI against the sources above · 29 Sept 2026. Spotted an error? Report it.