Technical

Srinivasa Reddy Kandi: Anthropic Unveils “Mythos” AI Model in Cybersecurity-Focused Project Glasswing

April, 08, 2026-04:19

Share: Facebook | Twitter | Whatsapp | Linkedin | Visits: 37682 | :2821

---

Anthropic Unveils “Mythos” AI Model in Cybersecurity-Focused Project Glasswing:

Anthropic has introduced a preview of its advanced AI model, Mythos, positioning it as one of its most powerful systems to date. The model is being rolled out to a select group of organizations as part of a new cybersecurity initiative called Project Glasswing.

Rather than a full public release, Mythos is being tested by a small set of partners focused on strengthening software security. Although it wasn’t specifically designed for cybersecurity, the model is being used to analyze both proprietary and open-source systems to detect vulnerabilities in code.

According to Anthropic, Mythos has already uncovered thousands of zero-day vulnerabilities in recent weeks, many of which are considered critical. Interestingly, a significant number of these flaws date back 10 to 20 years, highlighting long-standing security gaps.

Mythos is a general-purpose model within Anthropic’s Claude ecosystem, built with strong reasoning and coding capabilities. As a frontier model, it represents the company’s most advanced tier, aimed at handling complex tasks such as autonomous coding and agent-based workflows.

The initiative includes major partners like Amazon, Apple, Microsoft, Cisco, Broadcom, CrowdStrike, Linux Foundation, and Palo Alto Networks. These collaborators will share insights from their testing to help improve security practices across the industry. In addition to these partners, around 40 other organizations will gain limited access to the preview.

Anthropic also noted that it has been in discussions with U.S. federal authorities regarding the use of Mythos. However, those conversations may be complicated by an ongoing dispute between the company and government entities, following concerns raised by the Pentagon about supply-chain risks tied to Anthropic’s stance on limiting certain surveillance and targeting applications.

The existence of Mythos was first revealed through a data leak reported by Fortune, where an early draft document—referring to the model as “Capybara”—was accidentally exposed in a publicly accessible data repository. Anthropic later attributed the incident to human error. The leaked material described Mythos as a new class of AI model, surpassing the capabilities of its previous Opus series and representing a significant leap in performance.

Author: Kandi Srinivasa Reddy, Srinivasa Reddy Kandi, #KandiSrinivasaReddy, #SrinivasaReddyKandi