🛡️Introducing Shieldstral, Mistral’s 3B open-weights model for content safety that can be deployed on-device 🧵 http://mistral.ai/news/shie…

Mistral AI introduced Shieldstral, a 3‑billion‑parameter, open‑weights model designed for content‑safety tasks and capable of on‑device deployment. The announcement was shared via a tweet from @MistralAI with a link to the full release notes (https://mistral.ai/news/shieldstr…). Shieldstral aims to provide developers with an autonomous safety filter that can run locally without requiring cloud infrastructure.

Source: Mistral AI (X) | 2026-08-04