OpenAI has just announced Daybreak, a new cybersecurity platform applying artificial intelligence (AI), designed to help businesses detect software vulnerabilities, identify cyber threats and automatically fix security incidents faster.
This is considered OpenAI's strongest step forward in the field of AI network defense.
The appearance of Daybreak takes place in the context of AI deeply changing the cybersecurity industry. Instead of relying mainly on humans as before, many systems can now automatically detect attacks, analyze risks and respond almost in real time.
OpenAI's move also came just one month after Anthropic introduced the Glasswing project and the Claude Mythos security model. This shows that major AI companies are expanding their competition into the field of cybersecurity, instead of just focusing on chatbots or creating content.
According to OpenAI, Daybreak is built as a continuous software security platform. The system combines GPT-5.5 models specializing in cybersecurity with Codex Security, an AI tool in the form of agents capable of searching, analyzing and patching vulnerabilities directly in real source code stores.
One of the notable points of Daybreak is the approach of bringing security right from the beginning of the software development process. Instead of waiting for the system to be attacked or a vulnerability to appear before fixing it, this platform helps businesses detect and handle security risks while writing and checking the source code.
OpenAI said the platform can support security experts in checking source code, building threat models, verifying patches, analyzing dependent libraries and prioritizing the most dangerous vulnerabilities.
Daybreak works by creating an attack risk simulation model based on enterprise source code. Then, the system will test possible attack scenarios to detect high-risk vulnerabilities.
Meanwhile, Codex Security plays the role of an automated operating class. This tool not only responds to commands but can also access the source code store, create patches, check bug fixes in a separate environment and send security reports for corporate audits.
OpenAI provides Daybreak with three different AI levels. Standard GPT-5.5 version for routine software development tasks. GPT-5.5 Trusted Access for Cybersecurity version is used for verified cybersecurity tasks such as malware analysis and vulnerability detection.
Meanwhile, the most advanced version GPT-5.5-Cyber supports in-depth tasks such as intrusion testing or attack simulation, but comes with a stricter control and verification process.
OpenAI said the platform was developed from the previous GPT-5.4-Cyber, which was a system that previously supported handling more than 3,000 security vulnerabilities. The company is currently partnering with many major names in the cybersecurity industry such as Cloudflare, Cisco, CrowdStrike, Oracle, Fortinet and Palo Alto Networks to integrate Daybreak into business systems.
