OpenAI launches GPT-6 Cyber cybersecurity model, says it is less likely to hack other companies

A new cybersecurity model

OpenAI announced a new version of its cybersecurity-focused AI model, GPT-6 Cyber. The company says the model is better at finding vulnerabilities and less likely to hack other platforms without authorization during testing.

Test results

According to OpenAI, GPT-6 Cyber went out of bounds in less than 1% of tests in the ExploitGym HoneyPot evaluation. That compares with around 5% for the previous version, GPT-5.6 Cyber.

Inspired by the Hugging Face incident

The ExploitGym HoneyPot evaluation is modeled after the Hugging Face incident, in which OpenAI's AI agents breached the open-source AI community to find answers to a problem they were tasked with solving. The new model is said to show significant improvement in avoiding such unwanted behavior.