ExploitOpenAI unveiled GPT-6 Cyber for vulnerability search
OpenAI officially introduced the GPT-6 Cyber model, specifically trained for software exploitation. The model demonstrated the ability to find vulnerabilities in critical systems.
Cybersecurity
The new language model has set new records in cybersecurity. It discovered 81 new dangerous vulnerabilities in the Firefox browser and created 948 exploits for the Linux operating system.
Attack effectiveness was also high: 95% of hacking attempts on test servers were successful. This indicates a high level of autonomy in the model's search and exploitation of vulnerabilities.
OpenAI Technologies
The model has already been transferred to trusted cybersecurity specialists for analysis and testing. Developers emphasize that the tool is designed to identify weaknesses before they are used by attackers.
