GPT-6 Astra: Frontier intelligence for work, now generally available in Microsoft Foundry
The future of enterprise AI isn’t just about chat interfaces; it’s about how effectively a model collaborates with you. Introducing GPT-6 Astra—OpenAI’s latest model, now available for all Microsoft Foundry users. This tool is designed to assist organisations in making informed decisions for complicated tasks and executing them seamlessly across the applications and systems they rely on.
We’re committed to enhancing your experience with advanced technologies like Astra. Many AI initiatives stumble when it comes to identity management, networking, governance, data processing, evaluation, and compliance. Microsoft Foundry integrates these essential elements within Azure, enabling teams to move swiftly from concept to execution with confidence.
Transform Open-Ended Goals into Tangible Actions
Astra is designed to tackle open-ended challenges, working through them step by step to devise a plan and deliver polished outcomes. It can assess various trade-offs, adapt to new directions as projects evolve, and utilize tools across different applications and systems.
For businesses, this means moving beyond simple conversational AI to produce significant deliverables:
- Strategic planning and decision support: Astra has the ability to deconstruct challenges into manageable steps, assess different options, present its recommendations, and highlight next steps for review.
- High-quality, goal-oriented outputs: With its contextual awareness and adherence to templates and standards, Astra can help generate well-crafted documents, spreadsheets, presentations, and analyses ready for expert scrutiny.
- Multifunctional execution: Astra can perform tasks within various applications on behalf of users, navigating multiple software environments and handling multi-step projects with the necessary human oversight.
At Replit, we aim to make valuable intelligence accessible to all. With GPT-6 Astra available via Microsoft Foundry, we unlock a new realm of capabilities that extend beyond just generating code to actively creating software. We’re thrilled about the potential for developers and entrepreneurs to create innovative applications with this intelligent software partner.
—Luis Hector Chavez, CTO, Replit
Application Use and Cross-Functionality
Astra is equipped to function across well-known applications, including workflows lacking dedicated APIs. It can read information displayed on screens and interact with approved interfaces, assisting with tasks like updating records, navigating development tools, testing software, and compiling results into comprehensive reports. OpenAI has achieved top-notch outcomes in specific evaluation tasks, although performance may vary based on task complexity, tools, configurations, and safeguards.
Such direct capabilities require careful management. Information displayed in applications may sometimes be incomplete, misleading, or intentionally designed to sway an agent’s reactions. Foundry aids clients in defining access rights, approval processes, and monitoring systems, while creating workflows that include scoped credentials, approved resources, and human check-ins for significant actions.
Notable Enterprise Applications We’ve Observed
- Software Development: Astra can replicate intricate bugs, explore potential reasons, suggest fixes, and prepare changes for developer assessment and testing.
- Data Analysis: Astra aids in crafting and improving dashboards within Power BI, enabling analysts to compare data, identify trade-offs, and prepare insights for distribution.
- Professional Documentation: Astra can generate documents, spreadsheets, and presentations that adhere to predefined templates and business standards, resulting in polished materials for expert evaluation.
- Operational Workflows: Astra can facilitate tasks such as updating customer information, processing forms, testing websites, and working through approved interfaces where dedicated APIs may not be available.
Enterprise Control Measures for Independent Tasks
OpenAI refers to Astra as its most closely aligned model yet and will release supporting materials on alignment, safety, and computational evaluations. Foundry enhances these features with robust enterprise security, safety, and compliance mechanisms, including Microsoft Entra identity and access management, data encryption, private networking, role-based access controls, content filtering, safety evaluations, and robust governance tools.
Importantly, prompts and outputs are not used for training these models. While these features enable customers to set up safeguards and maintain oversight, they do not completely eliminate risk or substitute each organisation’s duty to choose and configure controls suitable for their specific scenarios and regulatory needs.
At Albertsons Companies, we see the true benefit of cutting-edge AI as the capacity to adapt as rapidly as the technologies evolve, all while maintaining enterprise integrity. This involves fostering an environment that allows us to assess new features, implement the ones that provide value quickly, and uphold rigorous security, governance, and operational controls as we scale. Azure OpenAI within Microsoft Foundry allows us to strike a balance between speed and control, keeping our teams focused on delivering meaningful outcomes for our customers, employees, and the organisation.
—Anirban Nandi, VP, Data and AI, Albertsons Companies
Global Reach with Appropriate Service Levels
GPT-6 Astra is accessible in Foundry models with both Standard and Provisioned Throughput deployment options, available in both Global and US Data Zone regions. The Standard option offers pay-as-you-go flexibility for unpredictable needs, while the Provisioned Throughput option delivers dedicated model-processing power for tasks needing consistent latency and guaranteed throughput. Customers can select the most suitable deployment based on their workload demands.
Astra has been designed to be token-efficient, especially when handling complex tasks, helping clients manage their consumption as they grow. Actual usage and costs will depend on the workload and configuration settings.
Pricing for GPT-6 Astra*
| Deployment | Context Length | Pricing (USD $/million tokens) | |||
| Input | Cached Input | Cached Writes | Output | ||
| Standard Global | Short context | $10.00 | $1.00 | $12.50 | $50.00 |
| Long context | $20.00 | $2.00 | $25.00 | $75.00 | |
| Standard Data Zone (US) | Short context | $11.00 | $1.10 | $13.75 | $55.00 |
| Long context | $22.00 | $2.20 | $27.50 | $82.50 | |
Get Started Today
Discover the Model: Experience GPT-6 Astra in Foundry Models to see its advanced features in action.
Create Agentic Workflows: Begin with the Foundry Agent Service to introduce cross-application task execution into your processes.
Frequently Asked Questions
- What is GPT-6 Astra? GPT-6 Astra is OpenAI’s latest enterprise AI model designed to assist with complex decision-making and task execution across various applications.
- How can Astra improve business operations? Astra can break down challenges, provide insights, and automate workflows, thus enhancing efficiency and productivity.
- What are the applications of GPT-6 Astra? Astra is applicable in software development, data analysis, document production, and operational workflows.
- Is there a cost associated with using Astra? Yes, the pricing for using GPT-6 Astra varies by deployment type and context length. Detailed pricing can be found on the Azure OpenAI pricing page.
Share this content:
Discover more from Qureshi
Subscribe to get the latest posts sent to your email.