Microsoft added assist for the lately launched DeepSeek-R1 synthetic intelligence (AI) mannequin to its Azure AI Foundry and GitHub on Wednesday. The reasoning-focused AI mannequin can now be accessed through these platforms. The Redmond-based tech big additionally highlighted that it has performed rigorous security evaluations to make sure that the mannequin is secure to make use of for each particular person and industrial functions. Moreover, the corporate can be bringing distilled variations of the R1 mannequin to Copilot+ PC customers through the AI Toolkit.
Microsoft Brings DeepSeek-R1 to Azure AI Foundry, GitHub
In a weblog publish, the tech big introduced that the DeepSeek-R1 AI mannequin is now out there within the mannequin catalogue of Azure AI Foundry and GitHub. Notably, Azure AI Foundry is an enterprise-focused platform the place builders can construct, consider, and develop generative AI functions and customized copilots.
Making the announcement, Microsoft’s Company Vice President of AI Platform, Asha Sharma stated, “One of many key benefits of utilizing DeepSeek R1 or another mannequin on Azure AI Foundry is the velocity at which builders can experiment, iterate, and combine AI into their workflows.”
The publish additionally highlighted that the corporate performed crimson teaming and security evaluations on the R1 mannequin, together with automated evaluation of mannequin behaviour in addition to in depth safety evaluations. Microsoft acknowledged that it has ensured to mitigate potential dangers. Azure AI Foundry additionally supplies extra security through its Content material Security filtering system and the Security Analysis System.
To make use of DeepSeek-R1, customers can seek for it within the mannequin catalogue. After recognizing it, they must open the mannequin card and click on on deploy. This can present the consumer with the inference utility programming interface (API) required to run the mannequin.
Distilled DeepSeek-R1 Coming to Copilot+ PCs Customers
In a separate publish, Microsoft introduced that it’s bringing optimised variations of DeepSeek-R1 to Copilot+ PCs. These distilled variations will first arrive on the Snapdragon X-powered gadgets and later shall be added to the Intel Core Extremely 200V chipset and others. The primary mannequin launch would be the DeepSeek-R1-Distill-Qwen-1.5B mannequin which may be accessed through the AI Toolkit. The 7B and 14B variants shall be added later.