Anthropic’s Claude Mythos outperformed OpenAI’s GPT5.5 on real‑world Google Chrome vulnerability exploits, a new benchmark designed to test the performance of frontier AI models to exploit real-world vulnerabilities found . During Infosecurity Europe 2026, Bugcrowd presented the first findings of ExploitBench,… Read More "Mythos Outperforms GPT5.5 on Google Chrome Vulnerability Exploits"