OpenAI's GPT-5.5 Matches Claude Mythos in Identifying Security Vulnerabilities

The UK’s AI Security Institute evaluated OpenAI’s GPT-5.5 and found its ability to identify security vulnerabilities comparable to Claude Mythos. Both models were assessed for cybersecurity capabilities, with GPT-5.5 noted as generally available to the public. The Institute also analyzed a smaller, less expensive model that, despite requiring additional prompter scaffolding, demonstrated similar effectiveness. No specific technical metrics, dates, or vulnerability identifiers (e.g., CVEs) were provided in the evaluation. The findings highlight the growing role of AI in vulnerability detection without detailing operational impacts or deployment contexts.