OpenAI has launched GPT-6 Astra, its latest flagship artificial intelligence model, with major improvements in reasoning, coding, research, and computer use.
The company introduced the new model after years of work across AI training and alignment. Moreover, OpenAI describes Astra as its most intelligent and aligned model so far.
The new model focuses heavily on complex professional work. As a result, users can give it demanding tasks that require multiple steps, tools, and careful decision-making.
OpenAI says GPT-6 Astra can handle computer use, web browsing, software engineering, cybersecurity, science, and professional workflows at an advanced level.
In addition, the model can work across several stages of a task instead of simply answering individual questions. This capability could make it useful for businesses, researchers, developers and other professionals.
The company has also improved Astra’s ability to work directly with computers. For example, it can complete online forms, organize calendars and update customer information in business software.
Moreover, Astra can conduct online research and prepare useful summaries. It can also analyze scientific information, create plots, build websites, and check whether website features work correctly.
Another major focus involves professional document creation. GPT-6 Astra can produce documents, presentations, spreadsheets, and analyses while following existing templates and instructions.
This feature could help businesses maintain consistent formats across their work. At the same time, Astra can adapt when users change requirements mid-task.
OpenAI has also highlighted major improvements in software engineering. Developers can use the model for difficult coding tasks and workflows that require several connected steps.
Furthermore, the company has trained Astra to understand context more effectively. The model can make reasonable decisions when instructions leave small details unclear.
However, Astra can also ask focused questions when missing information could significantly change the final result. As a result, users can maintain greater control over important decisions.
OpenAI also reports impressive benchmark results for GPT-6 Astra. The model achieved a 98% score on FrontierMath Tier 4, according to the company.
OpenAI says Astra has already contributed to work involving long-standing mathematical problems. As a result, the model could become particularly useful for advanced scientific and mathematical research.
The company also reports a 99.9% score on ARC-AGI-3. Meanwhile, Astra achieved a 100% score on ExploitBench in OpenAI’s published results.
These benchmark scores demonstrate the model’s strength across different types of difficult tasks. However, real-world performance can still vary by task, available tools, and user instructions.
Computer use represents another important improvement. OpenAI says Astra establishes a new performance frontier in several computer and browser-based tasks.
The model can interact with digital tools while following a user’s broader objective. Consequently, it could reduce the manual work required for repetitive professional processes.
GPT-6 Astra also brings stronger visual judgment to digital creation. OpenAI says the model can create websites, games, applications, and other visual experiences from user instructions.
Moreover, Astra can perform quality checks after creating a website. This lets the model identify problems and improve its work over a longer workflow.
OpenAI has also focused heavily on safety and alignment. The company says Astra includes stronger safeguards for advanced computer and cybersecurity capabilities.
The company has introduced additional monitoring to detect situations where an AI agent may misunderstand instructions or move beyond its intended task. In some cases, safety systems can stop work and allow a user to review the situation.
Cybersecurity is especially important because Astra offers significantly stronger capabilities than earlier models. OpenAI says the model has reached the Critical cybersecurity capability level under its Preparedness Framework.
As a result, the company has strengthened protections against potentially harmful cyber actions. OpenAI has also introduced tighter internal security measures for developing and deploying advanced models.
The launch of GPT-6 Astra could also change how professionals use AI in everyday work. Instead of using AI only for writing or answering questions, users can increasingly delegate complete workflows.
For instance, a professional could ask Astra to research information, analyze the findings, and prepare a polished document. Similarly, developers could use it for coding, testing, and other software tasks.
The model also supports very large amounts of context. OpenAI’s developer documentation lists a 1.05-million-token context window and a maximum output of 128,000 tokens.
Meanwhile, OpenAI has rolled it out gradually rather than giving every user immediate access. The company initially opened Astra to a limited number of organizations.
OpenAI plans to expand availability to ChatGPT Plus, Pro, Business, and Enterprise users. Developers and organizations can also access Astra through the OpenAI API, Microsoft Azure, and AWS Bedrock.
For developers, OpenAI uses gpt-6-astra as the API model identifier. The company currently lists standard API pricing at $10 per million input tokens and $50 per million output tokens.
The arrival of GPT-6 Astra marks another major step in OpenAI’s push toward AI systems that can complete complex work rather than generate individual responses.
As access expands, businesses and professionals will get a clearer picture of how much Astra can improve everyday workflows. Its combination of reasoning, computer use, research, and content creation could significantly expand AI’s role in professional environments.
