OpenAI released AI security agent tool Codex Security, challenging the traditional cybersecurity industry.

OpenAI is launching an artificial intelligence agent aimed at helping security teams identify and fix vulnerabilities in large databases, which may reduce the need for traditional cybersecurity companies. In a statement on Friday, OpenAI said that the intelligent agent named Codex Security is able to identify cybersecurity vulnerabilities and propose solutions, and then proceed to fix these vulnerabilities. The company stated that the tool is designed to operate at a "large scale" and provide "easily deployable patches", allowing developers to focus on higher-level tasks. The company also revealed that the intelligent agent tool has been used to scan and identify security vulnerabilities in open source code repositories. The product developed by ChatGPT competes with Claude Code Security from Anthropic PBC, which can also discover vulnerabilities and provide solutions. After Claude released this new security tool last month, cybersecurity stocks saw a significant decline. At that time, the stock prices of Crowdstrike (CRWD.US) and Cloudflare (NET.US) both dropped by 8%. Codex Security evolved from Aardvark, which is an intelligent cybersecurity research tool based on GPT-5 that OpenAI released last year for beta testing with a small group of clients. Codex Security will be released in the form of a research preview version for ChatGPT enterprise, commercial, and academic clients in the coming days, with the first month free. The intelligent agent is designed to learn from feedback in order to improve its threat model over time.