OpenAI has launched GPT-6 Astra, its latest flagship artificial intelligence model, with expanded capabilities across computer use, software development, research, cybersecurity and professional tasks.
The company describes Astra as its most intelligent and aligned model to date. OpenAI President Greg Brockman called the release a “generational leap” and suggested that it could eventually be viewed as a milestone in the development of artificial general intelligence. OpenAI, however, has not formally declared that Astra constitutes AGI.
Astra is designed to move beyond answering prompts by completing complex, multi-step tasks on computers. OpenAI says the model can browse websites, work with software, conduct research and create documents, spreadsheets and presentations. Its API supports a context window of 1.05 million tokens and up to 128,000 output tokens.
The model is initially being rolled out to a limited set of organisations, with broader availability planned over the coming days for ChatGPT Plus, Pro, Business and Enterprise users. Developers will also be able to access Astra through the OpenAI API, Microsoft Azure and Amazon Bedrock. OpenAI has priced standard API access at $10 per million input tokens and $50 per million output tokens.
Cybersecurity is a significant part of the release. Astra is OpenAI's first model to reach the “Critical” cybersecurity capability level under the company's Preparedness Framework. OpenAI says the model can, when provided with appropriate tools and access, identify previously unknown vulnerabilities and develop ways to exploit weaknesses in protected systems without requiring human guidance at every step.
During internal evaluations, Astra identified two previously unknown vulnerabilities, which OpenAI said it disclosed to the relevant software maintainers. The company has introduced additional safeguards because the same capabilities that can help security teams identify weaknesses could potentially be misused for offensive purposes.
OpenAI is also positioning Astra as a model capable of carrying out longer end-to-end workflows. The company says it has improved performance across browsing, coding, science and professional work, reflecting the industry's broader move from conversational AI towards systems capable of independently executing tasks.
The launch comes as competition among developers of frontier AI models intensifies and companies increasingly focus on agents that can interact directly with digital environments. Astra's rollout will now test how OpenAI balances greater autonomy and computer-use capabilities with the safety and monitoring requirements accompanying more capable AI systems.
Disclaimer: This article may include information derived from interviews, press releases, public statements, research, company communications and other publicly available or third-party sources. Such material may be summarised, paraphrased or contextualised for journalistic and editorial purposes. All rights in third-party content remain with their respective owners.