OpenAI Brings GPT-5.4-Cyber, AI That Stops Cyberattacks Before They Happen

Tech giant OpenAI has introduced its new AI model, GPT-5.4-Cyber. This model is built to support cybersecurity professionals, as it focuses on detecting and fixing critical security vulnerabilities faster. Notably, OpenAI’s rollout of GPT-5.4-Cyber comes shortly after its rival Anthropic recently announced its Mythos AI, which has not been made available to individuals due to safety concerns. What Is GPT-5.4-Cyber?GPT-5.4-Cyber is the latest AI model from OpenAI, trained to assist users with defensive cybersecurity tasks. Unlike general AI models, this one is specially designed to understand security workflows and respond more effectively to identify security bugs and analyse the risks in software. The Sam Altman-led firm claims that the model is ‘cyber-ermissive,’

meaning it is less likely to block legitimate security-related queries. “Customers in the highest tiers will get access to GPT‑5.4‑Cyber, a model purposely fine-tuned for additional cyber capabilities and with fewer capability restrictions. This is a version of GPT‑5.4, which lowers the refusal boundary for legitimate cybersecurity work and enables new capabilities for advanced defensive workflows, including binary reverse engineering capabilities that enable security professionals to analyze compiled software for malware potential, vulnerabilities, and security robustness without needing access to its source code,” wrote OpenAI in its blog post published on Tuesday.

Who Can Use GPT‑5.4‑Cyber Works?The American tech company asserted that the cybersecurity tool is dual-sense. In simple words, this tool can be used for protection and exploitation. Due to this, the company has restricted its access to selected users. Under OpenAI’s Trusted Access for Cyber (TAC) program, individuals applying for this AI will be identified first. Moreover, organisations can request access for the teams.

OpenAI notes that integrating the GPT‑5.4‑Cyber into software development will help the experts to fix issues before they occur, getting real-time feedback and immediately patching the vulnerabilities as they build. Meanwhile, Anthropic also announced its Mythos model, bringing ‘Project Glasswing’, which works similarly to detect software vulnerabilities and fix them. OpenAI’s GPT‑5.4‑Cyber will emerge as a strong competitor to Mythos.