OpenAI CEO Sam Altman has announced a new cybersecurity initiative called “Daybreak”. This new AI model, developed by OpenAI, is being touted as a Claude Mythos rival by tech experts, as it aims to help companies identify and fix software vulnerabilities using artificial intelligence.
In a post on ‘X’, Sam Altman wrote: “OpenAI is launching Daybreak, our effort to accelerate cyber defence and continuously secure software.”
According to OpenAI, Daybreak is designed to improve cyber defence by finding risks in software systems before hackers can exploit them. OpenAI said the initiative combines its latest GPT-5.5 models with Codex, the company’s coding-focused AI system, to support developers and security teams.
What is Daybreak, and what can it do?
OpenAI said Daybreak can assist with several cybersecurity tasks, including secure code reviews, threat modelling, patch validation, malware analysis, and identifying risky software dependencies. The company claims the system can help reduce the time needed to detect and fix vulnerabilities in software.
Importantly, the company has listed three model options under Daybreak. GPT-5.5 is the default model with standard safeguards for general-purpose, developer and knowledge work. GPT-5.5 with Trusted Access for Cyber is meant for verified defensive work in authorised environments. This includes secure code review, vulnerability triage, malware analysis, detection engineering and patch validation.
The initiative also includes “Trusted Access for Cyber”, a special framework that gives verified cybersecurity professionals access to advanced AI tools with fewer restrictions. OpenAI stated that stronger identity verification and security controls will be used for sensitive cyber-related tasks.
OpenAI CEO Sam Altman said AI is becoming increasingly capable in cybersecurity and can play an important role in helping organisations protect their systems from attacks.
Competition in AI cybersecurity intensifies
The launch of Daybreak highlights growing competition among AI companies in the cybersecurity sector. Anthropic’s Claude Mythos recently gained attention after reports claimed it identified hundreds of vulnerabilities in major software systems. However, Anthropic has not released Mythos publicly due to concerns over its powerful cyber capabilities.
Industry experts believe AI-powered cybersecurity tools could become increasingly important as companies face rising digital threats. OpenAI said Daybreak aims to make software “resilient by design” by integrating security into the development process from the beginning instead of addressing vulnerabilities only after attacks occur.
Several major technology and cybersecurity firms, including Cisco, Intel, SentinelOne and Snyk, are expected to work with OpenAI to evaluate the new platform.