OpenAI Unveils GPT-5.6 Sol with Enhanced Cybersecurity Features

In a significant development, OpenAI has released a limited preview of its GPT-5. 6 models, including Sol, Terra, and Luna, to a select group of companies and the U. S. government.

The flagship Sol model boasts the most robust safety stack to date, with strengthened protections against high-risk activity, sensitive cyber requests, and repeated misuse. This enhanced security feature makes it an ideal tool for vulnerability research and exploitation. OpenAI has pressure-tested the system to identify weaknesses and hardened it against real-world attacks. The model's capabilities are designed to facilitate legitimate work such as code review, vulnerability research, and patch development, while enforcing strong safeguards to prevent offensive activity.

The company aims to ensure that these advanced capabilities benefit defenders, enabling them to identify weaknesses, develop patches, and strengthen systems. Although the preview phase may involve some teething issues, with users potentially encountering blocked or paused requests, OpenAI is working to refine the system. The GPT-5. 6 Sol model has demonstrated impressive results in finding vulnerabilities in code and developing exploits, and OpenAI plans to make the models generally available in the coming weeks.

This release follows the company's recent launch of an improved GPT-5. 5-Cyber model and its collaboration with Trail of Bits on the Patch the Planet project to secure open-source projects.

Source: The Hackers News