Artificial intelligence is now being used to make software more secure.

Anthropic announced a new initiative called Project Glasswing, which Apple is also a partner in. As part of Glasswing, Anthropic is sharing a preview of its newly introduced Claude Mythos model with selected partners, including Apple.

Anthropic states that Mythos has found "thousands of high-priority vulnerabilities" in "every major operating system and web browser."

Apple is Part of Anthropic's Project Glasswing Cybersecurity Initiative

Apple is among the top technology companies forming Anthropic's Project Glasswing group.

Today we are announcing Project Glasswing; a new initiative that brings together Amazon Web Services, Anthropic, Apple, Broadcom, Cisco, CrowdStrike, Google, JPMorganChase, the Linux Foundation, Microsoft, NVIDIA, and Palo Alto Networks. Our goal is to secure the world's most critical software.

Additionally, Anthropic mentions that more than 40 additional organizations that "develop or maintain critical software" have access to the Mythos Preview AI model. These software organizations are expected to use Mythos to discover and fix vulnerabilities before the AI model is released to the world.

Claude Mythos Found Thousands of Bugs in Every Major Operating System and Browser

Claude Mythos is reported to be used to find serious security vulnerabilities in every major operating system and web browser.

Mythos Preview has found thousands of high-priority vulnerabilities in "every major operating system and web browser." Given the pace of AI advancement, it won't be long before such capabilities spread beyond determined actors applying them safely. The implications for economies, public safety, and national security could be serious. Project Glasswing is an urgent initiative to use these capabilities for defense purposes.

The company states that these vulnerabilities were able to be identified based on "decades of human reviews and millions of automated security tests."

One example is finding and combining a Linux kernel security vulnerability that could provide full control over a machine.

Cybersecurity expertise is just one of the strengths of the new Claude Mythos AI model. Anthropic's latest model shows advancements in reasoning, agency search, and computer usage, especially in agency coding compared to Claude Opus 4.6.

Anthropic has published a system card detailing the latest metrics for Claude Mythos Preview.

Accessibility

“We do not plan to make Claude Mythos Preview generally accessible,” says Anthropic, “but our ultimate goal is to enable our users to safely scale Mythos-class models; this is for cybersecurity purposes, but also for many other benefits that such high-capability models will bring.”

To learn more about Project Glasswing and Claude Mythos, you can find Anthropic's announcement here.

Do More with Your iPhone

Apple AirTag 2 | Add Find My tracking to anything

Beats USB-A to USB-C Cable | Official CarPlay cable

Wireless CarPlay Adapter | No need to plug in anymore

AirPods 4 | Apple's most affordable ANC AirPods