OpenAI rolls out more capable version of its cybersecurity model OpenAI is releasing an updated and more capable version of its GPT-5.5-Cyber model, designed for advanced and authorized cybersecurity work, allowing for deeper code analysis and vulnerability validation. The company is also expanding access through initiatives like the OpenAI Daybreak Cyber Partner Program and supporting the Patch the Planet initiative to aid open-source maintainers. These advancements occur as policymakers scrutinize AI deployment, presenting a challenge for developers to balance powerful capabilities with preventing malicious use.

• OpenAI updates its GPT-5.5-Cyber model to be more permissive and capable for authorized cybersecurity tasks.
• The updated model enhances analysis of large codebases, identifies vulnerabilities, and aids in developing software patches.
• OpenAI is launching the Daybreak Cyber Partner Program for vetted companies to use GPT-5.5 for customer security.
• The company is also funding the Patch the Planet initiative to help open-source maintainers manage vulnerabilities.
• OpenAI has established partnerships with several countries and is working with critical infrastructure operators and government networks. Continue reading https://www.axios.com/2026/06/22/openai-rolls-out-more-capable-version-of-cyber-model