- Project Glasswing and Anthropic's Claude Mythos Preview Are Designed to Address AI Security Threats
- Mythos is so powerful that it will not be released to the public, only to select companies.
- The work has already found decades-old bugs and critical flaws in major operating systems and browsers.
Anthropic has lifted the wraps on Project Glasswing, a new cybersecurity initiative it leads alongside AWS, Anthropic, Apple, Broadcom, Cisco, CrowdStrike, Google, JPMorganChase, Linux Foundation, Microsoft, Nvidia, and Palo Alto Networks.
Glasswing is designed to identify and fix vulnerabilities in critical software using Claude Mythos Preview, which Anthropic describes as an “unreleased, general-purpose frontier model.”
In a nutshell, it marks the official use of AI by companies in the fight against AI itself: with AI-based cyber attacks increasing by the day, the coalition of companies is fighting fire with fire.
Article continues below.
Project Glasswing is here
“We formed Project Glasswing because of the capabilities we have observed in a new frontier model trained by Anthropic that we believe could reshape cybersecurity,” the company wrote in a blog post announcing the news.
Anthropic boasted that Mythos had already found thousands of high-severity vulnerabilities, identifying potential flaws in all major operating systems and web browsers, including decades-old bugs that had not been detected by humans. In addition to detecting bugs, Mythos can also generate exploits and propose or generate patches for a complete cycle.
However, Mythos is only being used cautiously by the approved group of companies because Anthropic considers it too powerful and risky for an open release. If misused, it could dramatically increase cyberattacks as it is capable of autonomously generating exploits.
Delivered through cloud providers such as AWS and Google Cloud, more than 40 other organizations that maintain critical software will also be able to gain access to the model.
Mythos consistently outperforms Claude's Opus 4.6 in agent coding, reasoning, and agent search/computer usage benchmarks, more than doubling Opus 4.6's performance in SWE-bench Multimodal.
Even with the proactive approach, Anthropic says companies alone are not responsible for managing the effects of AI on cybersecurity: “cutting-edge AI developers, other software companies, security researchers, open source maintainers, and governments around the world have essential roles to play.”
Follow TechRadar on Google News and add us as a preferred source to receive news, reviews and opinions from our experts in your feeds. Be sure to click the Follow button!
And of course you can also follow TechRadar on TikTok for news, reviews, unboxings in video form and receive regular updates from us on WhatsApp also.
