A Chinese cybersecurity firm has claimed AI-driven vulnerability discovery capabilities that approach the scale of those attributed to Anthropic’s recently unveiled Claude Mythos model. The claims have been analyzed by Eugenio Benincasa, an ETH Zurich cybersecurity researcher focusing on China, in a post published on the Natto Thoughts blog. Anthropic claims that its new Mythos frontier model has autonomously discovered thousands of vulnerabilities. To prevent potential abuse, Mythos has not been publicly released and is only available to a few dozen major organizations through Project Glasswing. However, Anthropic’s own chief executive has suggested that open source models and Chinese developers could replicate Mythos-level performance within 6-12 months, a view echoed by researchers at cloud security firm Wiz.

Report: China’s 360 Digital Security Group has uncovered ~1,000 previously unknown vulnerabilities, including in Microsoft’s Office, using an AI-powered agent.