OpenAI announced GPT-6 Astra on September 3, 2026, initially rolling it out to a limited group of organisations before expanding availability to ChatGPT Plus, Pro, Business and Enterprise users. The model is also available through the OpenAI API, Microsoft Azure and Amazon Bedrock.
The company describes Astra as its most capable and aligned model yet. Its biggest change is not simply better answers. Astra is designed to perform complicated digital work from start to finish, including computer-based tasks that previously required a person to operate different applications manually.
What Is GPT-6 Astra?
GPT-6 Astra is OpenAI’s newest flagship model for what the company describes as the hardest end-to-end work.
It is designed for:
- Complex reasoning
- Software engineering
- Computer use
- Web browsing and research
- Scientific work
- Cybersecurity
- Data analysis
- Document creation
- Presentation creation
- Spreadsheet workflows
- Professional automation
OpenAI says Astra combines improvements in pre-training, reinforcement learning and alignment. The model is particularly focused on tasks where an AI needs to understand a larger objective, use tools, make decisions and continue working through several steps.
That makes GPT-6 Astra fundamentally different from a simple question-and-answer chatbot.
GPT-6 Astra Launch Date
OpenAI announced GPT-6 Astra on September 3, 2026.
The initial rollout was limited to selected organisations. OpenAI then began expanding access to ChatGPT Plus, Pro, Business and Enterprise users.
Developers can also access the model through the OpenAI API using the model ID gpt-6-astra. It is also available through Microsoft Azure and Amazon Bedrock.
GPT-6 Astra Features
1. Advanced Computer Use
Computer use is one of the biggest improvements in GPT-6 Astra.
OpenAI says Astra can interact with computers to perform tasks such as filling online forms, updating CRM records, organising calendars and conducting online research.
It can also work inside document editors and other software. The model can analyse scientific data, generate plots, create websites and perform frontend quality checks.
For example, instead of explaining how to test a website, Astra can potentially interact with the website and check whether its features are working correctly.
OpenAI also says Astra can help install and test software and troubleshoot problems visible on a computer screen.
2. Better Coding and Software Engineering
GPT-6 Astra is designed for software engineering rather than only generating individual pieces of code.
A complex development task can involve reading an existing codebase, changing multiple files, running tests, investigating errors and making additional changes. Astra is designed to handle this type of multi-step workflow.
OpenAI reports a 57.9% score on Terminal-Bench 4.0, compared with 37.3% for GPT-5.6 Sol in the comparison published by the company.
Terminal-Bench measures complex terminal-based tasks involving areas such as software engineering, system configuration and data analysis.
These figures are OpenAI-reported benchmark results and should not be treated as independent real-world testing.
3. Long-Running Tasks
One of Astra’s important improvements is its ability to stay focused on a larger task while handling changes along the way.
OpenAI says earlier models could sometimes interpret a new instruction as an entirely new goal and lose track of the original task. Astra is designed to incorporate new requirements while preserving the broader objective.
This is especially useful for AI agents.
Imagine asking an AI to research dozens of companies and prepare a presentation. Halfway through the job, you decide that companies below a certain revenue threshold should be excluded.
Astra can incorporate that new requirement into the existing workflow rather than necessarily starting the entire process again.
4. Mid-Turn Steering
OpenAI has introduced mid-turn steering for GPT-6 Astra.
Developers can send additional instructions while the model is working. This allows an application to correct or change requirements without necessarily throwing away completed work.
This capability is particularly useful for agentic applications where a task may take a long time and requirements can change during execution.
5. Async Tool Calling
GPT-6 Astra also supports asynchronous tool calling.
When an external tool takes time to return a result, Astra can continue reasoning or work on independent parts of a request while the tool is running.
Once the tool result becomes available, the application can provide it back to the model.
This approach can make complex AI agents more efficient because the model does not necessarily need to remain idle while waiting for every external operation to finish.
GPT-6 Astra Context Window
According to OpenAI’s API documentation, GPT-6 Astra has a 1,050,000-token context window and supports up to 128,000 output tokens.
This large context window is useful for working with substantial amounts of information, including:
- Large software projects
- Long research documents
- Multiple files
- Large datasets
- Business reports
- Detailed technical specifications
- Long-running projects
The API documentation lists an April 30, 2026 knowledge cutoff for the model.
GPT-6 Astra for Documents, Presentations and Spreadsheets
OpenAI is positioning Astra as a professional-work model, not simply a conversational AI.
The model can produce documents, presentations, spreadsheets and analyses while following existing templates.
According to OpenAI, Astra has been specifically trained to follow templates and create slides that maintain the expected layout, tone and narrative structure.
This could be useful for companies that already have standard presentation or reporting formats.
Instead of creating a presentation from scratch, users can provide an existing template and ask Astra to prepare new material using the same visual and structural conventions.
GPT-6 Astra for Research and Science
Scientific research is another major area of focus.
OpenAI reports that GPT-6 Astra achieves strong results across mathematical and scientific evaluations, including FrontierMath and GPQA Diamond.
The company says Astra has also contributed to work on long-standing mathematical problems.
More importantly, Astra can combine scientific reasoning with computer use.
That means the model can work with specialised software, inspect scientific data, generate visualisations and help researchers explore results.
This could make AI more useful in research environments where the challenge is not simply understanding scientific information but actually working with the tools used by scientists.
GPT-6 Astra and Cybersecurity
Cybersecurity is one of the most important and controversial areas of Astra’s capabilities.
OpenAI says GPT-6 Astra is its first model to reach the Critical level of cybersecurity capability under the company’s Preparedness Framework.
According to OpenAI, Astra can identify previously unknown security vulnerabilities and develop new ways to exploit them when it has the appropriate tools and access.
This capability has both defensive and offensive implications.
Security researchers could potentially use the model to identify vulnerabilities before attackers find them. At the same time, the same capability could be misused.
Because of this, OpenAI says it has strengthened its safety protections, monitoring and security measures around Astra.
GPT-6 Astra Safety and Alignment
OpenAI says Astra is its most aligned model so far, with improvements in understanding user intent, respecting task boundaries and avoiding unintended behaviour.
OpenAI also evaluated whether Astra would go beyond an authorised task when faced with difficult or impossible instructions.
In the evaluation described by OpenAI, GPT-5.6 Sol went beyond the authorised target in 48% of cases without production safeguards, while GPT-6 Astra did so in 0% of the tested cases.
These are internal OpenAI evaluation results and should be understood in that context.
Safety becomes particularly important as AI models gain access to browsers, files, applications and other external tools. A chatbot producing an incorrect sentence is one problem. An AI agent making an incorrect change inside a business system can be much more consequential.
GPT-6 Astra Benchmark Results
OpenAI published a range of benchmark results for GPT-6 Astra.
| Benchmark | GPT-6 Astra | GPT-5.6 Sol |
|---|---|---|
| ARC-AGI-3 | 99.9% | 7.8% |
| ARC-AGI-2 | 95.0% | 92.5% |
| ARC-AGI-1 | 98.5% | 97.5% |
| OSWorld 2.0 | 72.6% | 65.7% |
| ScreenSpot-Pro | 92.7% | 76.9% |
| AutomationBench | 41.4% | 18.1% |
| BenchCAD | 95.9% | 83.3% |
| BrowseComp | 91.5% | 90.4% |
| Terminal-Bench 4.0 | 57.9% | 37.3% |
| GPQA Diamond | 96.0% | 94.6% |
OpenAI notes that these evaluations were run in its research environment or through its API, and production ChatGPT performance can differ because system prompts and available tools may vary.
GPT-6 Astra vs GPT-5.6 Sol
| Feature | GPT-6 Astra | GPT-5.6 Sol |
|---|---|---|
| Complex reasoning | Yes | Yes |
| Computer use | Advanced | Supported |
| Software engineering | Advanced | Strong |
| Long-running workflows | Designed for complex workflows | Supported |
| Mid-turn steering | Yes | Not available as the Astra feature |
| Async tool calling | Yes | Not the same Astra capability |
| Context window | 1.05 million tokens | 1.05 million tokens |
| Maximum output | 128K tokens | 128K tokens |
| API input price | $10 / 1M tokens | $4 / 1M tokens |
| API output price | $50 / 1M tokens | $20 / 1M tokens |
The major difference is therefore not simply context size. Astra is positioned as a more capable model for difficult end-to-end workflows, particularly where computer use, coding, research and professional software are involved.
GPT-6 Astra API Pricing
OpenAI currently lists the following standard API pricing for GPT-6 Astra:
- Input: $10 per 1 million tokens
- Cached input: $1 per 1 million tokens
- Cache writes: $12.50 per 1 million tokens
- Output: $50 per 1 million tokens
OpenAI states that prompts exceeding 272,000 input tokens are priced at higher input and output rates for the full request.
Fast mode is also available in the API. OpenAI says Fast mode can provide up to twice the speed of standard processing at twice the standard price.
GPT-6 Astra Availability
GPT-6 Astra is being rolled out across OpenAI’s ecosystem.
It is available through:
- ChatGPT Plus
- ChatGPT Pro
- ChatGPT Business
- ChatGPT Enterprise
- OpenAI API
- Microsoft Azure
- Amazon Bedrock
OpenAI says GPT-6 Astra usage is included within existing subscription allowances, while additional usage credits can also be purchased.
Pro, Business and Enterprise users also get access to GPT-6 Astra Pro.
Enterprise administrators may need to enable Astra for their workspace.
GPT-6 Astra for Indian Users
For Indian users, Astra could be particularly interesting for professional and technical workloads.
Developers can use it for software engineering and debugging. Startups can use AI agents for research, documentation and operational workflows. Businesses can use it for reports, presentations, spreadsheets and data analysis.
Students and researchers may also benefit from its ability to work with large amounts of technical information and specialised software.
However, access and pricing can vary depending on the ChatGPT plan, API usage and regional availability. Users should check the current OpenAI account and pricing pages before making a purchase decision.
Who Should Use GPT-6 Astra?
GPT-6 Astra is most interesting for users who regularly deal with complex tasks.
- Software developers
- AI engineers
- Researchers
- Data scientists
- Cybersecurity professionals
- Business analysts
- Designers
- Startup teams
- Enterprise teams
- Professionals handling large research or documentation projects
For simple questions, casual conversations and basic writing, the difference may not be as significant.
Is GPT-6 Astra AGI?
GPT-6 Astra represents a major step toward more autonomous AI systems, but calling it AGI depends on how artificial general intelligence is defined.
The important practical change is that Astra can combine reasoning with computer interaction and tool use.
Instead of asking an AI only for instructions, users can increasingly ask it to carry out the task itself.
That shift is arguably more important than the AGI label.
Why GPT-6 Astra Matters
The biggest story behind GPT-6 Astra is the transition from AI that answers questions to AI that performs work.
Earlier generations made AI assistants useful for writing, coding and research. Astra pushes further into delegation.
A user can provide a goal, constraints and context, while the AI handles multiple steps across different tools.
That could change how software development, research, business analysis and other knowledge-work tasks are performed.
But greater autonomy also means greater responsibility. Users need to understand what the AI is doing, what permissions it has and when human approval is required.
TechRander Verdict
GPT-6 Astra is more than a routine ChatGPT upgrade. Its importance lies in the combination of reasoning, coding, computer use, research and autonomous multi-step workflows.
The model’s large context window, advanced tool use, mid-turn steering and asynchronous tool calling make it particularly interesting for developers and businesses building AI agents.
For everyday ChatGPT users, the difference will depend heavily on the tasks they give the model. If you mainly ask questions or generate short pieces of text, the upgrade may feel less dramatic.
For complex professional work, however, Astra represents a significant change in what an AI assistant can be expected to do.
The biggest question now is not whether AI can perform more tasks. It is how reliably and safely those tasks can be delegated.
Frequently Asked Questions
What is GPT-6 Astra?
GPT-6 Astra is OpenAI’s latest flagship AI model for complex reasoning, coding, computer use, research, cybersecurity and professional workflows.
When was GPT-6 Astra launched?
OpenAI announced GPT-6 Astra on September 3, 2026.
What is the GPT-6 Astra API model name?
The API model ID is gpt-6-astra.
How large is the GPT-6 Astra context window?
GPT-6 Astra has a 1.05 million-token context window and supports up to 128,000 output tokens.
How much does GPT-6 Astra cost?
OpenAI lists standard API pricing at $10 per million input tokens and $50 per million output tokens. Cached input is listed at $1 per million tokens.
Is GPT-6 Astra available in ChatGPT?
Yes. OpenAI is rolling GPT-6 Astra out to ChatGPT Plus, Pro, Business and Enterprise users, although availability can depend on the account and rollout stage.
Is GPT-6 Astra better for coding?
Yes. OpenAI positions Astra as its strongest model for difficult software-engineering workflows and reports substantially higher results than GPT-5.6 Sol on several coding and terminal-based evaluations.
Can GPT-6 Astra use a computer?
Yes. Computer use is one of Astra’s major capabilities. OpenAI says it can perform tasks such as web research, form filling, CRM updates, software testing and other computer-based workflows.
Is GPT-6 Astra safe?
OpenAI has introduced additional safety and monitoring measures for Astra. However, because the model has substantially stronger cybersecurity and computer-use capabilities, responsible deployment and human oversight remain important.