OpenAI Expands Daybreak: GPT-5.5-Cyber & IBM Join

What did OpenAI announce with Daybreak on June 22, 2026?

OpenAI expanded its Daybreak cybersecurity initiative on June 22, 2026, releasing the full version of GPT-5.5-Cyber, an updated Codex Security plugin, a new Cyber Partner Program, and a program called Patch the Planet. The company described the push as an effort to patch vulnerable software at machine speed. IBM joined as one of the program's launch partners.

How does GPT-5.5-Cyber compare to GPT-5.5?

GPT-5.5-Cyber is a specialized model built for verified defenders who need advanced cyber capabilities paired with verification, monitoring, and scoped controls. It posts clear gains over the base GPT-5.5 model across three benchmarks.

The ExploitGym jump — from 25.95% to 39.5% — is the largest relative gain across the three tests, according to Investing.com.

What is the Codex Security plugin and what has it done so far?

Codex Security is a plugin that integrates into OpenAI's Codex to find, validate, and fix software vulnerabilities. Since its March research preview, it has scanned over 30 million commits across more than 30,000 codebases. Human reviewers have manually marked more than 70,000 findings as fixed. Over 500,000 findings have been automatically determined to be fixed.

The plugin identifies vulnerabilities, checks code reachability, gathers validation evidence, develops patches, and verifies results — all within a single workflow.

Who is in the OpenAI Daybreak Cyber Partner Program?

OpenAI launched the Daybreak Cyber Partner Program with a group of security providers. Partners can use GPT-5.5 with Trusted Access for Cyber inside their own security products and services. Launch partners include:

You might also like

Agility Digit: Live at Amazon, GXO & Toyota

Nvidia Banned Chips Double in Price on China

CAISI Signs AI Reviews With Google, Microsoft

Meta Smart Glasses Switch to Muse Spark AI

• Accenture
• Akamai
• Check Point
• Cisco
• Cloudflare
• CrowdStrike
• IBM
• Palo Alto Networks

OpenAI said it will expand the program to more organizations in coming months. For builders tracking AI security partnerships, this is one of the broadest enterprise rollouts of a frontier AI model into active defense workflows to date.

What is IBM's specific role in Daybreak?

IBM joined the Daybreak Cyber Partner Program and launched a new application security service on the same day. The service uses OpenAI's cyber model capabilities to help organizations identify and validate software vulnerabilities faster. It goes past traditional code scanning by using AI-driven analysis to assess application code and prioritize areas most likely to contain flaws and exploitable paths.

The service runs on IBM Consulting Advantage, IBM's AI platform for delivering consulting services. It operates inside the client's environment with read-only access to code repositories and bounded execution. Clients can start with focused evaluations and expand to continuous monitoring over time.

IBM's Global Managing Partner for Cybersecurity Services, Mark Hughes, said: "Attackers are already using AI to probe, exploit, and scale threats at machine speed. Defenders need the same advantage, with the security and control enterprises require."

OpenAI CISO Dane Stuckey added: "Security is central to realizing the benefits of advanced AI."

IBM's involvement also connects to Project Lightwell, an enterprise security clearinghouse backed by a $5 billion commitment from IBM and Red Hat. As IBM's newsroom reports, Lightwell uses OpenAI's cyber capabilities alongside other frontier AI models for code review and remediation.

What is Patch the Planet?

Patch the Planet is an initiative co-founded by OpenAI with Trail of Bits, in collaboration with HackerOne and Calif. It funds security researchers to work directly with open-source maintainers. More than 30 open-source projects have committed to participate. Initial participants include:

• cURL
• Go
• Python
• Sigstore
• pyca/cryptography

Trail of Bits has engineers working full-time across 19 open-source projects using Codex and GPT-5.5-Cyber. They have already identified hundreds of security issues and merged dozens of patches, with more still under coordinated disclosure.

The model scanned security-relevant components across more than 30 million lines of code. It generated proof-of-concepts for 8 kernel pointer information leaks and 24 local privilege escalation vulnerabilities in the Linux Kernel.

Which governments have Trusted Access for Cyber partnerships with OpenAI?

OpenAI has established Trusted Access for Cyber partnerships with Australia, Canada, France, Germany, Japan, the Republic of Korea, and EU institutions including ENISA.

Here's what we know so far: these are government-level access agreements, not just commercial deals — a distinction that signals OpenAI is positioning GPT-5.5-Cyber as infrastructure for national-level defense workflows, not only enterprise software.

The new IBM application security service is available today. OpenAI said it will expand the Daybreak Cyber Partner Program to more organizations in coming months. For context on how AI is being applied to active enterprise threats, see our coverage of AI-driven security exploits.