OpenAI Releases New GPT-6 Astra Frontier Model

OpenAI released GPT -6 Astra yesterday, its latest frontier model that comes nearly two months after the general availability of the GPT-5.6 family of models. The ChatGPT developer described GPT-6 as “the world’s most intelligent and aligned model,” but it’s launching first for select organizations.

GPT-6 Astra will become available for all ChatGPT Plus, Pro, Business, and Enterprise users in the coming days. Developers will also be able to access it via the OpenAI API and AWS. Last month, OpenAI said it would reassess how it develops new frontier models after one of its internal research models went rogue and hacked the AI platform Hugging Face. As a result, GPT-6 includes additional safeguards to prevent “unintended consequences.”

“Astra excels at exercising care, respecting task boundaries, and communicating transparently. This work is the latest product of our long-running research program focused on training models that remain aligned with human intent from start to finish,” OpenAI said yesterday.

OpenAI mentioned different benchmarks to highlight that GPT-6 Astra delivers “significant efficiency gains in real knowledge-work tasks,” including game development, electrical engineering, and everyday knowledge work. The new frontier model should also excel at solving complex problems, running multi-step tasks, and creating polished documents that respect templates.

Regarding cybersecurity, OpenAI emphasized that GPT-6 Astra obtained a perfect 100% score on the ExploitBench benchmark, which analyzes the ability of an AI model to exploit vulnerabilities, while its previous GPT-5.6 model obtained a 78.5% score. “With the version of Astra launching today, defenders can use it to complete tasks such as secure code review and patching. However, Astra will refuse to comply with more advanced cybersecurity tasks such as creating proof-of-concept exploits for vulnerabilities,” OpenAI said about the additional security safeguards implemented in Astra.

“We believe it is the best model in the world for computer use, professional work, science, coding, cybersecurity, and more, OpenAI CEO Sam Altman said yesterday. “It took us some extra time to ensure that we could meet the safety and alignment standards required for this capability level, but we think you’ll find it worth the wait.”