Shieldstral

Shieldstral

Mistral AI Open and portable generative AI for devs and businesses. Mistral AI：面向开发者和企业的开放式便携生成式 AI。

AI Infrastructure Tools • Foundation Models

• We’re committed to empower the AI community with open technology. Our open models set the bar for efficiency, and are available for free, with fully permissive license.
• Our optimized commercial models are designed for performance and are available via our flexible deployment options.

AI 基础设施工具 • 基础模型

• 我们致力于通过开放技术赋能 AI 社区。我们的开源模型树立了效率标杆，并以完全宽松的许可协议免费提供。
• 我们经过优化的商业模型专为高性能设计，并可通过灵活的部署选项获取。

Overview This is the 23rd launch from Mistral AI. 概述 这是 Mistral AI 的第 23 次发布。

Shieldstral Launching today: Define safety at runtime for text and images. Shieldstral 今日发布：在运行时为文本和图像定义安全性。

Shieldstral is a 3B open-weight multimodal guardrail from Mistral. Define safety policies in natural language at inference time. It evaluates text, images, or both from a single token output, running locally on a single 16GB GPU.

Shieldstral 是 Mistral 推出的一个 3B 参数的开源权重多模态护栏模型。它允许在推理时通过自然语言定义安全策略。该模型通过单个 Token 输出即可评估文本、图像或两者，并可在单张 16GB 显存的 GPU 上本地运行。

Launch tags: Open Source • Artificial Intelligence • Security 发布标签：开源 • 人工智能 • 安全