← 回總覽

Anthropic 推出用于网络安全的 Claude Mythos 和 Project Glasswing

📅 2026-04-08 03:55 Boris Cherny 人工智能 3 分鐘 3750 字 評分: 92
Claude Mythos Project Glasswing Anthropic 网络安全 AI 安全
📌 一句话摘要 Anthropic 研究员 Boris Cherny 介绍了 Claude Mythos Preview 的发布,这是一款专门用于发现软件漏洞的前沿模型,也是新发起的 Project Glasswing 计划的一部分。 📝 详细摘要 Anthropic 的 Boris Cherny 讨论了“Claude Mythos Preview”的发布,这是一款专为网络安全设计的专业前沿模型。作为“Project Glasswing”计划的核心,该模型在识别软件漏洞方面表现出卓越的能力,超越了绝大多数人类专家。Cherny 强调了 Anthropic 负责任的发布策略,即优先向网络安全

Title: Anthropic Introduces Claude Mythos and Project Glasswing ...

URL Source: https://www.bestblogs.dev/status/2041605852382351666

Published Time: 2026-04-07 19:55:41

Markdown Content: Skip to main content ![Image 1: LogoBestBlogs](https://www.bestblogs.dev/ "BestBlogs.dev")Toggle navigation menu Toggle navigation menuArticlesPodcastsVideosTweetsSourcesNewsletters

⌘K

Change language Switch ThemeSign In

Narrow Mode

Anthropic Introduces Claude Mythos and Project Glasswing for Cybersecurity

Anthropic Introduces Claude Mythos and Project Glasswing for Cybersecurity

![Image 2: Boris Cherny](https://www.bestblogs.dev/en/tweets?sourceId=SOURCE_a68249cb) ### Boris Cherny

@bcherny

Mythos is very powerful, and should feel terrifying. I am proud of our approach to responsibly preview it with cyber defenders, rather than generally releasing it into the wild.

Model card here: www-cdn.anthropic.com/53566bf5440a10…

!Image 3: Anthropic

#### Anthropic

@AnthropicAI · 4h ago

Introducing Project Glasswing: an urgent initiative to help secure the world’s most critical software.

It’s powered by our newest frontier model, Claude Mythos Preview, which can find software vulnerabilities better than all but the most skilled humans. anthropic.com/glasswingShow More

783

2,243

16.1K

5.3M From

Apr 7, 2026, 7:55 PM View on X

244 Replies

190 Retweets

3,337 Likes

235.2K Views ![Image 4: Boris Cherny](https://www.bestblogs.dev/en/tweets?sourceid=a68249cb) Boris Cherny @bcherny

One Sentence Summary

Anthropic researcher Boris Cherny highlights the release of Claude Mythos Preview, a frontier model specialized in finding software vulnerabilities as part of the new Project Glasswing initiative.

Summary

Boris Cherny from Anthropic discusses the launch of 'Claude Mythos Preview', a specialized frontier model designed for cybersecurity. Integrated into 'Project Glasswing', the model demonstrates superior capabilities in identifying software vulnerabilities, surpassing most human experts. Cherny emphasizes Anthropic's responsible release strategy, which involves previewing the tool with cyber defenders rather than a general public release, and provides the official model card for technical reference.

AI Score

92

Influence Score 504

Published At Today

Language

English

Tags

Claude Mythos

Project Glasswing

Anthropic

Cybersecurity

AI Safety HomeArticlesPodcastsVideosTweets

Anthropic Introduces Claude Mythos and Project Glasswing ...

查看原文 → 發佈: 2026-04-08 03:55:41 收錄: 2026-04-08 06:00:32

🤖 問 AI

針對這篇文章提問,AI 會根據文章內容回答。按 Ctrl+Enter 送出。