OpenAI is rolling out GPT-6 Astra, its latest flagship AI model, with access expanding gradually across ChatGPT plans and products.
The company announced Astra on September 3, describing it as its most capable model for complex reasoning, coding, research, computer use and longer multi-step tasks. OpenAI initially made the model available to a limited set of organizations, with broader access planned over the following days.
For ChatGPT Plus subscribers, the rollout is currently focused on ChatGPT Work and Codex. OpenAI’s latest help documentation says Plus users will receive GPT-6 Astra in Work and Codex as access expands, while availability can differ between Chat, Work and Codex during the rollout. That means a Plus subscriber may see Astra in Work or Codex before it appears elsewhere in ChatGPT.
OpenAI says GPT-6 Astra is being positioned as a model for demanding end-to-end work rather than simple conversational tasks. It is designed to handle computer use, web browsing, coding, scientific work, cybersecurity, research and professional workflows, including creating documents, spreadsheets and presentations while adapting to changing instructions.
One of Astra’s biggest changes is its ability to work with computers and complete longer sequences of actions. OpenAI says it can perform tasks such as filling out online forms, updating customer records, organizing calendars, conducting online research, drafting content in documents or email, analyzing scientific data and even creating websites and checking whether their front-end features work correctly.
The model is also significantly stronger in cybersecurity-related testing. OpenAI reports that Astra scored 100% on its ExploitBench evaluation, compared with 78.5% for GPT-5.6 Sol, and achieved a 42.4% success rate on ExploitGym compared with 30.3% for the previous model. OpenAI also evaluated Astra on vulnerabilities disclosed between June and August 2026 and reported substantially higher arbitrary-code-execution rates than GPT-5.6 Sol.
Those capabilities also explain why safety has been a major part of the Astra rollout. OpenAI says Astra is its first model to reach the Critical level of cybersecurity capability under its Preparedness Framework. The company has added additional safety monitoring designed to detect cases where agents may not have interpreted instructions correctly, with conversations potentially being paused or stopped for review.
OpenAI also says Astra is included within existing subscription allowances rather than requiring a separate subscription. Plus and Business Standard users have limited Astra usage in Work and Codex, with additional usage available through credits. OpenAI notes that Astra can consume usage faster than GPT-5.6 Sol depending on the task, input and output size, reasoning settings and mode.
For developers, GPT-6 Astra is available through the OpenAI API as gpt-6-astra, with the company also making it available through Microsoft Azure and Amazon Bedrock. OpenAI lists a standard API price of $10 per million input tokens and $50 per million output tokens.
Access is still changing as the rollout continues. OpenAI’s current documentation says Pro users receive GPT-6 Astra in Chat, Work and Codex as the rollout progresses, while Plus users receive it in Work and Codex. Enterprise access can also depend on workspace model-access permissions.
For now, the key takeaway is that GPT-6 Astra is not arriving everywhere at the same time. Users may see it in one ChatGPT experience while it remains unavailable in another, and OpenAI is continuing to expand access. The company has positioned Astra as a major step toward AI systems that can take on longer, more complex tasks with less step-by-step guidance.
If this article helped you, please consider supporting our work. Every small contribution keeps Abijita.com independent and running.
GPT-6 Astra Is Rolling Out to ChatGPT Plus, With Broader Access Still Expanding
