OpenAI Codex Security Cloud
OpenAI announced an upgrade to Codex Security Cloud at DevDay on 29 September 2026, saying it includes access to cyber-capable models through Daybreak Blue by default, scans entire GitHub repos and continuously reviews new commits.
Tool Overview
Features, steps and FAQ below
Features
- ✓ OpenAI says Codex Security Cloud can scan entire GitHub repositories in the cloud and continuously review code on a schedule or whenever new commits arrive
- ✓ OpenAI says it investigates and deduplicates findings and prepares fixes for review in the cloud, even when your laptop is closed
- ✓ OpenAI says the upgraded Codex Security Cloud includes access to cyber-capable models through Daybreak Blue by default; OpenAI describes Daybreak Blue as an alias for its flagship general-purpose models with safeguards calibrated for defensive cybersecurity work
- ✓ OpenAI says Codex Security is a research preview for ChatGPT Enterprise, Edu, Business and Pro users that connects to GitHub repositories, builds a threat model, validates vulnerabilities in isolated environments and proposes fixes
- ✓ OpenAI says it is available as a plugin in Codex desktop and web
How to Use
- Use a ChatGPT Enterprise, Edu, Business or Pro plan and enable Codex Cloud and Codex Security access in workspace permissions
- Connect your GitHub repository through the Codex Security plugin in Codex desktop or web
- OpenAI says you can run a full-repository or per-commit scan; Codex validates vulnerabilities in isolated environments and generates fixes for review
- For higher-capability defensive security work, request Daybreak access through Trusted Access for Cyber following OpenAI's guidance
FAQ
What is Codex Security Cloud?
OpenAI says it is a managed application-security capability for connected GitHub repositories that scans code in the cloud on a schedule or on new commits, investigates and deduplicates findings, and prepares fixes.
Which models does it use by default?
OpenAI says the upgraded Codex Security Cloud includes access to cyber-capable models through Daybreak Blue by default; OpenAI describes Daybreak Blue as an alias for its flagship general-purpose models with safeguards calibrated for defensive cybersecurity work.
Who can use it?
OpenAI says Codex Security is a research preview for ChatGPT Enterprise, Edu, Business and Pro users, and requires Codex Cloud and Codex Security access to be enabled in workspace permissions.
How does it work?
OpenAI says it connects to GitHub repositories and builds a codebase-specific threat model, can scan a full repository or commits, validates potential vulnerabilities in isolated environments and proposes fixes for teams to review.
Where can I use it?
OpenAI says it is available as a plugin in Codex desktop and web, and runs in the cloud even when your laptop is closed.