The ChatGPT-maker’s product rivals Anthropic PBC’s Claude Code Security, which additionally spots vulnerabilities and suggests options to fix them.
| Photograph Credit score:
Dado Ruvic
OpenAI is introducing an AI agent that’s meant to assist security groups discover and patch vulnerabilities in giant databases, probably reducing into demand for legacy cyber companies.
The agent, known as Codex Security, works by figuring out cybersecurity flaws and proposing options earlier than fixing the bugs, OpenAI stated in an announcement Friday. The tool is designed to “function at scale” and supply “easy-to-accept patches,” and thus allow builders to deal with higher-level duties, the corporate stated. The San Francisco-based agency additionally shared that the agentic tool has been used to scan and establish security vulnerabilities inside open-source repositories.
The ChatGPT-maker’s product rivals Anthropic PBC’s Claude Code Security, which additionally spots vulnerabilities and suggests options to fix them. The discharge of Caude’s new security tool final month despatched cybersecurity shares tumbling. Each Crowdstrike Holdings and Cloudflare Inc fell by 8 per cent on the time.
Codex Security advanced from Aardvark, an agentic security researcher powered by GPT‑5 which OpenAI launched final yr to beta take a look at on a small cohort of consumers. Codex Security will likely be rolled out within the following days as a research preview to ChatGPT Enterprise, Enterprise and school clients with the primary month being free.
The agent is designed to study from suggestions to enhance its risk fashions over time.
Extra tales like this can be found on bloomberg.com
Revealed on March 7, 2026
Source link
#OpenAI #releases #agent #security #tool #research #preview
