OpenAI (OPENAI) has released a private beta version of Aardvark, a security research agent that autonomously monitors code to identify and help fix vulnerabilities in software.
“Aardvark represents a new defender-first model: an agentic security researcher that partners with teams by delivering continuous protection as code evolves,” OpenAI said. “By catching vulnerabilities early, validating real-world exploitability, and offering clear fixes, Aardvark can strengthen security without slowing innovation.”
It is powered with GPT-5 and is integrated with OpenAI Codex to help correct vulnerabilities by attaching a Codex-generated and Aardvark-scanned patch to each finding for review and one-click patching. Although it is primarily built for security, it can also detect bugs such as logic flaws and incomplete fixes.
Microsoft-backed (MSFT) OpenAI has already been utilizing Aardvark internally for several months across its own codebases and with its closest partners. It has also been applied to open-source projects.
The private beta version is now available to select partners. Interested parties can apply here.