Microsoft recognised for the second consecutive year as a Leader in the 2025 Gartner® Magic Quadrant for Data Science and Machine Learning Platforms

We’re delighted to announce that Microsoft has once again been recognised as a Leader in the 2025 Gartner® Magic Quadrant for Data Science and Machine Learning (DSML) Platforms. This accolade reaffirms our dedication to offering organisations an extensive toolkit for designing and deploying machine learning models and AI applications, completely transforming business operations. Azure Machine Learning seamlessly integrates with the broader, interoperable ecosystem that includes Microsoft Fabric, Microsoft Purview, and Azure AI Foundry.

Gartner defines a data science and machine learning platform as a consolidated set of code-based libraries paired with low-code tools. Such platforms are designed to facilitate independent work and collaboration among data scientists, business professionals, and IT teams. They provide automation and AI support throughout the data science lifecycle, which comprises phases like business understanding, data access and preparation, model creation, and sharing insights. These platforms also aid engineering workflows, involving data creation, feature engineering, deployment, and testing pipelines. They can be accessed via desktop applications or online browsers, with the option of using dedicated compute instances or through a fully managed cloud solution.

Setting the Standard in 2025

With Microsoft, we’re transforming our media expertise into a competitive edge by harnessing data to build brands and fuel business growth.

—Callum Anderson, Global Director for DevOps and SRE at Dentsu.

At Microsoft, our vision is a cohesive environment where data scientists, AI engineers, developers, IT operations professionals, and business users collaborate to create applications and manage the entire AI lifecycle across various personas and projects. Contributing to this vision, we launched Azure AI Foundry in November 2024. This platform allows developers to design, customise, and oversee AI applications. Azure Machine Learning operates atop Azure AI Foundry, providing technology for model customisation, fine-tuning, and RAG (Retrieval-Augmented Generation).

Pioneering AI with Azure Machine Learning and Intelligent Agents

The Foundry Agent Service, an integral part of Azure AI Foundry, enables development teams to coordinate AI agents that automate intricate, cross-departmental workflows. Whether creating solutions for software engineering, business process automation, customer support, or data analysis, the Foundry Agent Service offers a solid, secure, and interoperable platform for implementing AI agents in live environments.

• With features for multi-agent orchestration, developers have the flexibility to create agent systems that cooperate across tasks, share state information, recover from failures, and adapt as requirements change. These agents can be grounded in enterprise knowledge via Microsoft Fabric, Bing, and SharePoint, while also compatible with both proprietary and third-party tools through open standards such as MCP (Model Context Protocol) and A2A (Agent2Agent).

• Developers can kick off local projects using open-source frameworks like Semantic Kernel and AutoGen. We are steadily advancing towards offering a unified SDK across these two frameworks and Azure AI Foundry, allowing for a smooth transition from local experimentation to cloud production without the need for code rewrites. This ensures a consistent developer experience from initial prototyping to managed orchestration with observational insights and enterprise-grade control.

Together, Azure Machine Learning and the Foundry Agent Service create a future where AI systems are purpose-built for enterprise use, prioritising scalability and security.

Maximising AI Models using Azure AI Foundry

Azure AI Foundry presents developers with a cutting-edge way to deploy and manage over 11,000 AI models using tools like the Model Router, Model Leaderboard, and Model Benchmarks.

• The Model Leaderboard offers an efficient comparison of model performance across different real-world tasks. It provides transparent benchmark scores, task-specific rankings, and live updates, enabling users to select the model that best meets their needs in terms of accuracy, speed, and value.
• Model Benchmarks in Azure AI Foundry simplify the assessment of model performance using standard datasets and enable customers to evaluate models based on their own data to identify the best options for their specific needs.
• Additionally, the Model Router—now available for Azure OpenAI models—dynamically directs queries to the most appropriate large language model (LLM) based on factors like query complexity, cost, and performance, ensuring high-quality outcomes while keeping compute expenses low.

These features empower businesses to deploy flexible and adaptive AI systems with superior performance, security, and governance. By leveraging integrated innovation from Microsoft and its ecosystem, users gain access to forward-thinking solutions that enhance efficiency and scalability, helping them maintain a competitive edge in the rapidly changing AI landscape.

Enhancing AI Performance through Fine-Tuning in Azure AI Foundry

Fine-tuning is crucial for organisations looking to tailor pre-trained AI models to specific tasks, boosting their performance, accuracy, and adaptability while also lowering operational costs. Fine-tuning via Azure AI Foundry benefits from the underlying Azure Machine Learning tool chain.

• With innovations like Reinforcement Fine-Tuning (RFT) using the o4-mini model, Azure AI Foundry enables developers to enhance reasoning, context-aware responses, and dynamic decision-making through reinforcement signals. This adaptability is particularly valuable for applications requiring continuous learning, making it an ideal approach for evolving business logic to ensure models remain applicable in rapidly changing environments.

• Moreover, Azure AI Foundry streamlines fine-tuning with features like Global Training and the Developer Tier. Global Training reduces costs by allowing model customisation across multiple Azure regions, granting developers flexibility while maintaining strict privacy protocols. The Developer Tier offers an economical way to evaluate fine-tuned models, allowing for simultaneous testing across deployments and enabling users to accurately select the best candidates for production.

These capabilities collectively unlock the full potential of AI systems for developers and enterprises, fostering innovation and efficiency in today’s swiftly evolving digital landscape.

Empowering Organisations to Implement AI Solutions

From sectors like healthcare and finance to manufacturing and retail, customers leverage Azure Machine Learning to tackle complex challenges, optimise operations, and unlock new business models. Whether deploying foundation models, orchestrating AI agents, or scaling real-time inference, Microsoft is dedicated to assisting organisations in transforming data into impactful results.

Start Your Journey with Azure Machine Learning

The transition to Azure is just the beginning. We’ve established the groundwork to discover opportunities we once thought impossible.

—Steve Fortune, Chief Digital and Technology Officer at CSX.

Machine learning is fundamentally changing how businesses operate and compete in the digital era. It presents opportunities to enhance business processes, elevate customer experiences, and spur innovation. Azure Machine Learning stands as a robust and versatile platform for machine learning and data science, enabling organisations to responsibly and effectively implement AI solutions.

Gartner, Magic Quadrant for Data Science and Machine Learning Platforms, By Afraz Jaffri, Maryam Hassanlou, Tong Zhang, Deepak Seth, Yogesh Bhatt, 28 May 2025. 

GARTNER is a registered trademark and service mark of Gartner, Inc. and/or its affiliates in the U.S. and internationally. Magic Quadrant is a registered trademark of Gartner, Inc. and/or its affiliates and is used herein with permission. All rights reserved. 

This graphic was published by Gartner, Inc. as part of a larger research document and should be evaluated within the context of the entire document. The Gartner document is available upon request from [https://www.gartner.com/en/documents/6533902]. 

Gartner does not endorse any vendor, product, or service depicted in its research publications, and does not advise technology users to select only those vendors with the highest ratings or designations. Gartner research publications represent the opinions of Gartner’s research organisation and should not be interpreted as factual statements. Gartner disclaims all warranties, expressed or implied, regarding this research, including any warranties of merchantability or fitness for a particular purpose.