OpenAI Launches GPT-6 Astra With Major Advances

OpenAI has introduced GPT-6 Astra, a new frontier model with major improvements in computer use, software engineering, science, mathematics, cybersecurity, and professional work. In a range of specialized tests, Astra achieved near-perfect or perfect results, including evaluations of advanced mathematical reasoning, computer use, and cybersecurity.

Astra can handle complex multistep workflows, from conducting research and analyzing data to building websites, testing software, updating CRM records, and creating documents, spreadsheets, and presentations. It is also designed to follow existing templates, better understand user intent, and maintain context when tasks change or new instructions are added.

Astra expands AI’s capabilities for creating interactive digital content: from a text prompt, the model can create web apps and games. It can also work with complex 3D projects — for example, creating a model of a house and turning it into an interactive space that users can explore virtually before the house is built. 

OpenAI says GPT-6 Astra delivers major gains in software engineering, scientific reasoning, and cybersecurity. Its stronger ability to identify vulnerabilities also led the company to introduce additional safeguards and monitoring. The production version supports defensive security tasks but restricts more advanced exploit-development requests.

OpenAI is releasing GPT-6 Astra in waves, starting with a limited enterprise cohort before expanding access to ChatGPT Plus, Pro, Business, and Enterprise users. Developers will also be able to access it through the OpenAI API, Microsoft Azure, and AWS Bedrock. Standard pricing starts at $10 per million input tokens and $50 per million output tokens.