TLDR: Microsoft has announced significant advancements to its Azure AI Foundry platform, introducing the powerful GPT-5 model family, a new Browser Automation tool in public preview, and expanding the regional availability of its Agent Service. These updates aim to empower developers with more sophisticated generative AI capabilities, streamline application development, and enhance operational efficiency.
Microsoft is pushing the boundaries of generative AI development with its latest update to Azure AI Foundry, a managed platform designed for building, customizing, and deploying AI applications. The August and September 2025 announcements highlight the integration of the GPT-5 model family, the introduction of a Browser Automation tool, and an expanded footprint for its Agent Service, among other enhancements.
GPT-5 Family Integration:
The most prominent feature of this update is the availability of the GPT-5 model family within Azure AI Foundry. This includes `gpt-5`, `gpt-5-mini`, `gpt-5-nano`, and `gpt-5-chat`. The flagship `gpt-5` model is designed for complex, long-horizon tasks, supporting an impressive context window of up to approximately 272,000 tokens. The `gpt-5-chat` variant facilitates multimodal conversational use with around 128,000 tokens, enabling richer and more dynamic interactions. For real-time tool calling and ultra-low latency applications, the `mini` and `nano` models offer optimized performance.
Pricing for these models varies by deployment type. For global deployments, `gpt-5` is priced at $1.25 per million input tokens and $10.00 per million output tokens. The `gpt-5-nano` model offers a more cost-effective solution at $0.05 per million input tokens and $0.40 per million output tokens.
A key innovation with GPT-5 is ‘freeform tool calling.’ Microsoft states that this feature “enables the model to send raw text payloads like Python scripts, SQL queries, or configuration files directly to external tools without needing to wrap them in structured JSON—eliminating rigid schemas and reducing integration overhead.”
Introducing Browser Automation (Preview):
Azure AI Foundry now includes a Browser Automation tool, available in public preview. This tool allows AI agents to interact with real web browsers using natural language commands. It can handle a variety of tasks, such as navigation, form-filling, and appointment booking, automating last-mile UI steps that traditional APIs might miss. This capability is designed to operate within isolated, sandboxed sessions, ensuring privacy and security.
Expanded Agent Service Availability:
The Azure AI Foundry Agent Service has expanded its regional support, now available in four additional Azure regions: Brazil South, Germany West Central, Italy North, and South Central US. This brings the total number of supported regions to 17, enhancing accessibility and deployment options for developers globally.
Additional Model and Platform Enhancements:
Beyond the GPT-5 family, Microsoft has also introduced and updated other models. These include Black Forest Labs FLUX.1 Kontext [pro] and FLUX1.1 [pro], which are multimodal and text-to-image models offering faster generation and improved editing. Mistral Document AI has been added as a serverless OCR solution, capable of preserving layouts, tables, and figures while returning structured outputs.
Further platform advancements include new observability features that provide metrics on performance, quality, cost, and safety, all integrated into a detailed tracing dashboard. A new Model Router automatically selects the most suitable model for a given prompt, optimizing for quality and cost. The platform also features a Model Leaderboard to rank top-performing AI models across various categories and tasks.
Documentation Updates:
Microsoft has also updated its Azure AI Foundry documentation for September 2025, including new articles on the Azure AI Foundry status dashboard, Azure OpenAI in Azure AI Foundry Models v1 REST API reference, and detailed guides for the new Browser Automation tool.
Also Read:
- Zendesk Revolutionizes Customer Support with GPT-5 Integration, Achieving Significant Speed and Reliability Gains
- AWS Unveils Advanced AI Assistance in SageMaker Unified Studio with Agentic Chat, Q Developer CLI, and Multi-Cloud Integration
These comprehensive updates underscore Microsoft’s commitment to providing a robust and versatile platform for the next generation of AI applications, offering developers powerful tools and models to innovate and deploy at scale.
