DeepSeek R1 Launches on Azure AI Foundry and GitHub
DeepSeek R1 has officially joined the model catalog available on Azure AI Foundry and GitHub. This model is part of a vast array of over 1,800 models, which include various frontier, open-source, industry-specific, and task-oriented AI solutions. By being part of Azure AI Foundry, DeepSeek R1 provides businesses with a reliable, scalable, and enterprise-ready platform for integrating advanced AI into their operations.
With Microsoft’s commitment to reliability and innovation, DeepSeek R1 allows organizations to meet Service Level Agreements (SLAs), uphold security standards, and comply with responsible AI practices.
Accelerating AI Reasoning for Developers
The rapid accessibility of AI reasoning is reshaping how developers and companies utilize sophisticated intelligence. DeepSeek R1 presents a robust, cost-effective model, empowering a wider audience to benefit from advanced AI capabilities without a significant infrastructure outlay.
A standout advantage of DeepSeek R1, along with other models available on Azure AI Foundry, is the enhanced speed for developers to experiment, iterate, and integrate AI into existing workflows. Equipped with built-in model evaluation tools, developers can quickly compare outputs, benchmark performance, and scale their AI-driven applications. This level of accessibility, which was unimaginable just a short while ago, aligns with the vision of Azure AI Foundry to centralize outstanding AI models and accelerate innovation across enterprises globally.
Developing with Trustworthy AI
Our priority is to help customers build production-ready AI applications swiftly while upholding strong safety and security measures. DeepSeek R1 has undergone thorough evaluations, including red teaming and extensive safety assessments to ensure model behavior adheres to rigorous security standards. Additionally, Azure AI Content Safety provides built-in content filtering, allowing flexibility with opt-out options. Continuous automated evaluations within Azure AI Foundry enable clients to assess application outputs for quality and safety, both pre- and post-deployment. These safeguards ensure Azure AI Foundry remains a secure, compliant, and responsible environment for businesses to deploy AI solutions confidently.
Using DeepSeek in the Model Catalog
- If you don’t have an Azure subscription, you can sign up for an Azure account.
- Look for DeepSeek R1 in the model catalog.
- View the model card in the Azure AI Foundry model catalog.
- Click the deploy option to get the inference API key and also access the playground.
- The deployment page will show the API and key in under a minute, allowing you to try your prompts in the playground.
- You can utilize the API and key with various clients.
Get Started with DeepSeek R1 Today
DeepSeek R1 is now available via a serverless endpoint in the Azure AI Foundry model catalog. To get started, visit Azure AI Foundry and select the DeepSeek model.
On GitHub, you can find additional resources and detailed guides to help you integrate DeepSeek R1 into your applications seamlessly. You can also read the GitHub Models blog for more information.
Moreover, customers will have access to distilled versions of the DeepSeek R1 model, allowing them to run it locally on their Copilot+ PCs. Further details can be found in the Windows Developer blog.
As the model catalog in Azure AI Foundry continues to grow, there’s great anticipation to see how developers and businesses will use DeepSeek R1 to address real-life challenges and create transformative experiences. The aim is to offer a comprehensive portfolio of AI models, granting businesses of every size access to cutting-edge tools for fostering innovation and achieving success.
