OpenAI Launches GPT-6 Astra With Computer Use Tools and Broad Platform Rollout

OpenAI has unveiled GPT-6 Astra, their most advanced and aligned model to date. This new model focuses on computer use, software engineering, browsing, cybersecurity tasks, and professional knowledge work. Initially, Astra will be available to a select group of organizations, later expanding to paid ChatGPT users and developers through OpenAI's API, Microsoft Azure, and AWS Bedrock.

Astra surpasses earlier rumors about a special model launch, promising a staged multi-platform deployment with defined API pricing, large context capacity, and the ability to handle complex digital tasks. Businesses are particularly interested in whether Astra's computer-use functions can streamline manual work in existing processes.

The model is designed to work across tasks that typically require moving between software interfaces, web pages, and business documents. This expansion could help teams reduce repetitive administrative work. Astra also shows strong performance in code generation and professional knowledge work, including installing, testing, and troubleshooting software, as well as generating documents that follow a user's templates and style.

Key technical details include a context window of up to 1,050,000 tokens, API pricing of $10 per million input tokens and $50 per million output tokens, and a maximum output of 128,000 output tokens. There's also a Fast mode option that doubles speed at a higher price. Subscription access is included within existing allowances, with additional credits available for purchase.

Access to GPT-6 Astra is staged, starting with a limited set of organizations before being made available to all ChatGPT Plus, Pro, Business, and Enterprise users over the following days. The model will also be accessible through Microsoft Azure and AWS Bedrock. Organizations will need to enable Astra for their Pro, Business, and Enterprise workspaces, and developers should review their API access, quotas, and pricing configurations.

Astra's cybersecurity capabilities are rated at the Critical level under OpenAI's Preparedness Framework, indicating stronger protections against potential risks. This heightened security is crucial given the model's ability to browse, use software, and perform cybersecurity-related tasks.