Introducing Snowflake Arctic in Azure AI model catalog

We are excited for the next milestone in our expanded collaboration with Snowflake: hosting their innovative open source large language model (LLM), Snowflake Arctic, in Microsoft Azure . This collaboration underscores our commitment to pushing the boundaries of AI technology and providing our customers with state-of-the-art tools that drive innovation, efficiency, and transparency.



Expanding Collaboration in Large Language Models  

Arctic represents the pinnacle of open source AI technology with its unique Dense-Mixture-of-Experts (MoE) hybrid transformer architecture that activates 50% or fewer tokens than comparably sized models. This model is designed to deliver unparalleled efficiency and intelligence, making it ideally suited for the complex needs of modern enterprises. Arctic excels in a variety of tasks, including SQL code generation, sophisticated instruction following, and complex data analysis, positioning it as a leading tool for developers and businesses looking to leverage AI for real-world applications.

Snowflake has long been a valued partner of Microsoft. Our collaboration includes the introduction of Snowflake on Azure, a powerful integration that combines Snowflake's industry-leading data platform with Azure's robust capabilities enabling customers to manage diverse data sets seamlessly within a single system. Building on this foundation, we've co-developed a data connector that integrates Snowflake's platform with Azure . This joint development accelerates the entire model lifecycle, from data preparation to deployment, enabling faster and more efficient AI innovations.  

The integration of Arctic into the Azure AI model catalog not only expands our selection of models but also enhances our ability to meet diverse enterprise demands. Arctic is optimized for high performance across multiple benchmarks, setting new industry standards for both speed and resource efficiency.


Enhancing transparency in Enterprise AI  

Snowflake's decision to release Arctic under an Apache 2.0 license spurs innovation, allowing developers and researchers to frictionlessly use, fine-tune, and share model improvements back to the community. This open source innovation aligns with a core goal of Microsoft Azure: to promote an expansive development community for all developers both on and off of the Microsoft Azure platform.

On Azure, Arctic will be accessible alongside a vast array of AI models, giving users the freedom to choose and integrate the technologies and tools like prompt flow and Semantic Kernel that best fit their needs. This flexibility is crucial for building customized solutions that are both powerful and scalable.

“I'm excited about our partnership with Microsoft and the integration of Snowflake Arctic into Microsoft Azure,” said Baris Gultekin, Head of AI at Snowflake. “By making Arctic's state-of-the-art technology available on Azure AI model catalog, we are not only advancing the open source AI model landscape but also ensuring that diverse enterprise demands are met with the highest standards of efficiency and adaptability. This collaboration enables businesses to harness the full potential of AI, fostering an environment of growth and innovation across industries.”


Enhanced Security, Compliance, and Safety

As with all our offerings, the security and compliance of our users' data is paramount. Arctic's integration into Azure ensures that enterprises can deploy cutting-edge AI capabilities within the secure confines of Azure's globally recognized compliance and governance frameworks, all while benefiting from Arctic's high token efficiency and performance. This security-first approach ensures that our customers can innovate confidently, knowing their data is protected by rigorous security measures. In addition, developers can use Azure AI Content Safety to detect harmful user-generated and AI-generated content in applications, facilitating the development of safer AI applications.


A Future Built on Innovation and Openness

Looking forward, the addition of Snowflake Arctic to Azure AI is just the beginning. We are committed to continuously enhancing our AI offerings, ensuring that our customers have access to the most advanced and efficient models and tools in the market. This collaboration represents a significant step forward in our journey to build an AI-infused future where technology not only amplifies human potential but also fosters a culture of openness and collaboration across industries.


Getting Started with Arctic on Azure

We invite developers, data scientists, and businesses to explore the possibilities with Arctic in Azure AI Studio model catalog. To learn more about the new models visit Snowflake's press release and the announcement blog


This article was originally published by Microsoft's AI - Machine Learning Blog. You can find the original article here.