Anthropic launched Project Glasswing in partnership with Apple, Amazon, Google, Microsoft, and NVIDIA. The initiative provides partners with Claude Mythos Preview, an unreleased AI model capable of finding software vulnerabilities.
Anthropic claims the model outperforms most human experts in discovering security flaws. Claude Mythos has already identified thousands of vulnerabilities in major operating systems and browsers.
Partners will use the AI to secure critical infrastructure before malicious actors can exploit the technology. The announcement follows recent security discussions between United States officials and technology CEOs.
Anthropic is committing $100 million in usage credits to partners and 40 additional organizations. These credits will fund security improvements for both proprietary and open-source systems.