Anthropic Mythos AI: Public Rollout Soon for Vulnerability Detection

Anthropic is preparing to publicly roll out its Mythos-class AI models in the coming weeks. These advanced AI systems have already uncovered thousands of software vulnerabilities. What's interesting is that Anthropic confirmed these models will be available to customers very soon, following months of restricted testing. This announcement came with the launch of their upgraded Claude Opus 4.8 model. Mythos-class AI has gained attention for its ability to identify hidden software vulnerabilities at unprecedented speed. Anthropic states it can find thousands of critical security flaws overnight. So far, Mythos has been limited to preview access through Project Glasswing, involving major tech companies like Amazon Web Services and Google. This program aims to strengthen critical software infrastructure. Early testing shows significant results. Cloudflare, for example, discovered nearly 2,000 software bugs using Mythos Preview, with close to 400 being high-severity. Partners reportedly increased bug discovery rates by over ten times. The bottom line is these new AI systems could dramatically improve cybersecurity by finding vulnerabilities faster and more reliably.