Microsoft is proud to be named again as the leader of the 2025 Gartner® MagicQuadrant™ data science and machine learning (DSML) platform.
Microsoft is proud to be named again as the leader of the 2025 Gartner® MagicQuadrant™ data science and machine learning (DSML) platform. We believe this recognition reflects our ongoing commitment to providing organizations with a comprehensive toolchain for building and deploying machine learning models and AI applications, and transforming the way they operate their businesses. Azure Machine Learning is part of a wide interoperable ecosystem within Microsoft Fabric, Microsoft Purview, and Azure AI Foundry.
Gartner defines data science and machine learning platforms as integrated sets of code-based libraries and low-code tools. These platforms support the independent use and collaboration of data scientists and their businesses and IT counterparts. This includes automation and AI support throughout every stage of the data science lifecycle, including understanding business, accessing and preparing data, creating models, and sharing insights. It also supports engineering workflows, including data, features, deployment, and creating test pipelines. The platform is provided via a desktop client or browser with a supported computing instance, or as a fully managed cloud service.
Leading to 2025
With Microsoft, you transform your media expertise into a competitive advantage, leveraging data to build brands and drive business growth.
—Callum Anderson, global director of DevOps and SRE at Dentsu.
Microsoft envisions a unified experience where data scientists, AI engineers, developers, IT operations experts and business users gather together to create applications and manage the entire AI lifecycle of personas and projects. Therefore, in November 2024, we announced the availability of Azure AI Foundry, a platform that allows developers to design, customize and manage AI applications. Azure Machine Learning is a trusted workbench that resides on top of Azure AI Foundry, powered by the underlying toolchain technology, with the ability to tweak and model customization, including RAG.
Promote AI using Azure Machine Learning and Intelligent Agent
As part of Azure AI Foundry, Foundry Agent Service Empower developer teams to coordinate AI agents that automate complex, inter-duty workflows. Whether building solutions for software engineering, business process automation, customer support, or data analytics, Foundry Agent Service provides a robust, secure, and interoperable foundation for operating AI agents in production environments.
- Get support Multi-agent orchestrationdevelopers can design agent systems that coordinate across tasks, share state, recover from failures, and evolve flexibly as requirements change. These agents are based on enterprise knowledge using use Microsoft Fabric, Bing, and SharePointthanks to open standards like, while interacting with both your own and third-party tools. MCP (Model Context Protocol) and A2A (Agent2Agent).
- Developers can start building locally using open source frameworks like Semantic Kernel and Autogenand we Unified SDK Beyond two frameworks and Azure AI Foundry that allow you to move from local experiments to cloud production without rewriting your code. This ensures a consistent developer experience with managed orchestration and enterprise-grade control from initial prototyping.
together, Azure Machine Learning and Foundry Agent Service With scalability and security in mind, enable the future where AI systems are designed to use enterprises.
Use Azure AI Foundry to leverage AI models
Azure AI Foundry offers developers an innovative way to deploy and manage over 11,000 AI models using tools such as model routers, model leaderboards, and model benchmarks.
- The Model Leaderboard simplifies comparisons of model performance across real tasks, providing transparent benchmark scores, task-specific rankings, and live updates, allowing users to efficiently select high-precision, fast throughput, or competitive price performance ratios.
- Azure AI Foundry's model benchmarks provide a streamlined way to compare model performance using standardized datasets, allowing customers to evaluate models with their own data to identify the best fit for a particular scenario.
- Complementing this, model routers are currently available in Azure Openai models, and therefore evaluate factors such as query complexity, cost, and performance, placing queries on the most appropriate large language model (LLM) while ensuring high quality results and minimizing computing costs.
These capabilities allow businesses to deploy flexible and adaptable AI systems with enterprise-grade performance, security and governance. Integrated innovations from Microsoft and its ecosystem ensure users access future-ready solutions to improve efficiency and scalability, ensuring they are ahead of the rapidly evolving AI landscape.
Optimize AI performance with fine tuning in Azure AI Foundry
Fine tuning is an essential tool for organizations that aim to customize pre-trained AI models for specific tasks to improve performance, accuracy and adaptability. The Azure AI Foundry fine tuning comes with the underlying Azure Machine Learning Tool chain.
- With innovations such as reinforcement fine tuning (RFT) using the O4-MINI model, Azure AI Foundry allows developers to improve inference, context-conscious response and dynamic decision-making through enhanced signals. This adaptability is particularly suitable for applications that require continuous learning. This is an ideal way to evolve business logic and ensure that models are relevant in a dynamic environment.
- Azure AI Foundry further simplifies tweaks with features such as global training and developers. Global Training reduces costs by allowing model customization across multiple Azure regions, providing flexibility and scalability for developers while adhering to strict privacy policies. Developer Tier offers an affordable way to evaluate fine-tuned models, allowing simultaneous testing across deployments, allowing users to select the best candidate for production with accuracy and efficiency.
Together, these capabilities allow developers and businesses to unlock the full potential of AI systems and drive innovation and efficiency in the rapidly evolving digital landscape.
Allow your organization to deploy AI solutions
From healthcare and finance to manufacturing and retail, customers use Azure Machine Learning to solve complex problems, optimize operations, and unlock new business models. Whether it's deploying foundation models, tuning AI agents, scaling real-time inference, Microsoft helps organizations turn data into impact.
Start your journey with Azure Machine Learning
Migrating to Azure is just the beginning. We have laid the foundation for exploring opportunities that we could only imagine before.
– Steve Fortune, CSX Chief Digital and Technology Officer.
Machine learning is revolutionizing the operational and competitive landscape of businesses in the digital age. It offers opportunities to optimize business processes, improve customer experiences and drive innovation. Azure Machine Learning serves as a robust, versatile platform for machine learning and data science, enabling organizations to take responsibility and effectively implement AI solutions.
Gartner, The Magic Quadrant of Data Science and Machine Learning Platforms, Afraz Jaffri, Maryam Hassanlou, Tong Zhang, Deepak Seth, Yogesh Bhatt, May 28, 2025.
Gartner is Gartner, Inc. and/or registered trademarks and service marks of US and international affiliates, Magic Quadrant is Gartner, Inc. and/or its affiliates are registered trademarks and are used hereby with permission. Unauthorized reproduction is prohibited.
This graphic is by Gartner, Inc. should be published as part of a large research document by and evaluated in the context of the entire document. Gartner documentation is available upon request [https://www.gartner.com/en/documents/6533902].
Gartner does not endorse any vendors, products or services portrayed in research publications, and does not advise technology users to select only vendors with the highest ratings or other designations. Gartner's research publications are composed of the opinions of Gartner's research institutions and should not be interpreted as a statement of fact. Gartner disclaims all warranties, either expressed or implied, with regard to this research, including warranties of merchantability and fitness for a particular purpose.
