OpenAI Launches GPT-6 Astra with Enhanced Security

OpenAI announced on Thursday the rollout of its latest artificial intelligence model, GPT-6, also known as Astra, to select customers. OpenAI President Greg Brockman stated that safety is a top priority due to the model's advanced capabilities.

This release comes just over a year after the launch of GPT-5. OpenAI had previously paused model development for two weeks following a security breach involving two of its models at the AI platform Hugging Face.

Astra was developed with stronger safeguards, although it was not involved in the hack. The initial rollout will provide access to some cybersecurity customers, with a wider release for other paying customers to follow; users on free or lower-tier plans will not have access.

Astra is designed to autonomously handle various tasks, including website creation and scientific analysis, and can significantly reduce time spent on tasks like apartment hunting from six hours to under ten minutes. OpenAI's exclusivity agreement with Microsoft was terminated in April.