Anthropic, a competitor to OpenAI, has significantly expanded its Project Glasswing initiative, now allowing nearly 200 organizations worldwide to access its Claude Mythos Preview AI model, a substantial increase from around 50 organizations previously. This expansion is aimed at enhancing cybersecurity by enabling key industries to detect and address software vulnerabilities before they can be exploited by malicious actors. Organizations seeking access must comply with stringent security requirements. The announcement follows Anthropic's recent confidential IPO filing, which could lead to one of the largest tech listings in recent years. Since its launch in April, the Claude Mythos model has gained attention for its ability to identify and aggregate minor software flaws that could be exploited in cyberattacks, having already detected over 10,000 severe vulnerabilities. Anthropic aims to bolster software infrastructure ahead of the anticipated widespread availability of similar AI capabilities from other providers within the next six to 12 months.