On October 31, 2025, OpenAI introduced Aardvark, an AI-powered security agent that leverages the capabilities of GPT-5. This innovative agent is designed to autonomously identify and rectify security vulnerabilities within codebases. The project has now officially entered its internal testing phase.
