How Anthropic’s Mythos has rewritten Firefox’s approach to cybersecurity 07.05.2026

Mozilla's Firefox security team has significantly enhanced its bug detection capabilities through the use of Anthropic's advanced AI model, Mythos. Since implementing Mythos in April, researchers have discovered a substantial number of high-severity vulnerabilities, including some that were over a decade old, a marked improvement from previous AI tools that often generated low-quality reports. The agentic nature of Mythos allows it to assess its own findings, filtering out false positives and increasing efficiency. This has led to a dramatic increase in bug fixes, with 423 shipped in April 2026 compared to 31 the previous year, even uncovering complex sandbox vulnerabilities previously found infrequently by human researchers. Despite AI's progress in coding, human engineers still develop and review all patches.















