# Mistral 发布 3B 开源安全模型 Shieldstral

- 来源：🚨 AI News | TestingCatalog (@testingcatalog)
- 发布时间：2026-08-05 02:04
- AIHOT 分数：54
- AIHOT 链接：https://aihot.virxact.com/items/cmsez79wl1bsdro2em0ivvap6
- 原文链接：https://x.com/testingcatalog/status/2084702101607764270

## AI 摘要

Mistral AI 宣布推出 Shieldstral，一款用于内容安全的全新 3B 开源权重模型。

> 一款 3B 开源权重、策略自适应的多模态安全分类器，在文本安全方面可媲美最高 7 倍体量的模型，并在多模态审核方面树立了新标杆。

## 正文

Mistral AI announced Shieldstral， a new 3B open-weights model for content safety.

&gt； A 3B open-weight， policy-adaptive multimodal safety classifier that matches models up to 7x its size on text safety and sets a new state of the art on multimodal moderation.

### 引用推文

> Mistral AI：The model takes moderation policy as a plain-language question and returns a calibrated score. Text and images - one interface. Read the full technical report h...