OpenAI Launches Daybreak Cybersecurity Initiative

Jakarta (ANTARA) - OpenAI has launched the Daybreak cybersecurity initiative, designed to help companies detect, analyse, and remediate software vulnerabilities with the aid of artificial intelligence.

According to an Engadget report on Monday (11/5), Daybreak is developed on the premise that cyber defences should be built into software from the outset, rather than merely focusing on finding and fixing vulnerabilities.

According to the company, Daybreak is presented to prioritise the handling of high-impact issues and reduce analysis times from hours to minutes, in order to generate and test patches within repositories and deliver results with audit-ready evidence back to client systems.

OpenAI states that Daybreak uses the company’s artificial intelligence models, including the specialised agent Codex Security, which can be used to scan codebases, validate the highest-risk findings, and remediate them.

Daybreak will rely on the GPT-5.5 AI model for general purposes and GPT-5.5 with Trusted Access for Cyber for most defensive cybersecurity workflows, such as secure code reviews, malware analysis, patch validation, and threat detection engineering.

Meanwhile, the GPT-5.5-Cyber model is prepared for more specific needs, such as simulating cyberattacks, penetration testing, and controlled validation of security systems.

The Daybreak initiative is described as a competitor to Anthropic’s Project Glasswing, a cybersecurity service supported by the Claude Mythos Preview artificial intelligence model.