OpenAI Expands GPT-5.4-Cyber Access to Verified Defenders
Summary
OpenAI is expanding its Trusted Access for Cyber program to thousands of verified defenders and hundreds of teams, offering GPT-5.4-Cyber with relaxed guardrails and binary reverse-engineering capabilities. Model performance data was not disclosed.
Key points
- OpenAI is expanding its Trusted Access for Cyber program to thousands of verified defenders and hundreds of security teams.
- GPT-5.4-Cyber is a fine-tuned variant of GPT-5.4 with relaxed guardrails for legitimate cybersecurity work and binary reverse-engineering capabilities.
- The model is initially available on a limited, iterative basis to vetted security vendors, organizations, and researchers; individuals can apply through chatgpt[.]com/cyber.
- OpenAI has not shared performance information for GPT-5.4-Cyber.
- OpenAI said its Codex Security platform has helped identify more than 3,000 critical and high-severity vulnerabilities across the open-source ecosystem.
- The announcement follows Anthropic’s release of Claude Mythos, which is allegedly capable of autonomously discovering thousands of zero-day vulnerabilities; Anthropic has restricted access through Project Glasswing.
Article Details
- Event Type
- OpenAI expanded its Trusted Access for Cyber program and introduced GPT-5.4-Cyber, a cybersecurity-focused AI model for vetted defenders.
- Impact
- The program is scaling access to thousands of verified defenders and hundreds of security teams. GPT-5.4-Cyber includes binary reverse-engineering capabilities for analyzing compiled software for vulnerabilities and malicious behavior. Access is initially limited and iterative; OpenAI has not shared the model's performance. The article reports that Codex Security has helped identify more than 3,000 critical and high-severity vulnerabilities across the open-source ecosystem.
Vendors
AnthropicDays after Anthropic unveiled its Claude Mythos AI model, OpenAI introduced GPT-5.4-Cyber, a cybersecurity-focused model that will be offered to many defenders.OpenAIDays after Anthropic unveiled its Claude Mythos AI model, OpenAI introduced GPT-5.4-Cyber, a cybersecurity-focused model that will be offered to many defenders.
Products
Claude MythosDays after Anthropic unveiled its Claude Mythos AI model, OpenAI introduced GPT-5.4-Cyber, a cybersecurity-focused model that will be offered to many defenders.Codex Security(supporting the broader defender community through grants, open source contributions, and tools like Codex Security).GPT-5.4-CyberDays after Anthropic unveiled its Claude Mythos AI model, OpenAI introduced GPT-5.4-Cyber, a cybersecurity-focused model that will be offered to many defenders.Project GlasswingThis led Anthropic to withhold its public release and instead offer it only to a few dozen major organizations through a restricted program called Project Glasswing.