Microsoft continues to expand its cooperation at the artificial intelligence ecosystem. Although the company is known for its proximity to the chatgpt model of OpenAI, it also shows flexibility to support different artificial intelligence models. Finally, Microsoft announced that it will integrate the Deepseek R1 artificial intelligence model into Windows Copilot+ computers. In the first place, the model will be used in Snapdragon X processor devices, Intel Lunar Lake and AMD RYZEN AI 9 processors will come to computers in the future.
Deepseek-R1-Distill-QWEN-1.5B will be accessible for developers via Microsoft AI Toolkit. However, more powerful 7B and 14B variants are planned to be released later. Microsoft says that these models are optimized for Copilot+ devices and have the capacity to operate local.
Although the main version of the Deepseek R1 is less powerful than the larger 32B and 70B models, Microsoft emphasizes that these models are particularly optimized to work with NPU (neural processing unit). The minimum hardware requirements of Copilot+ Computers are determined as an NPU with at least 256 GB storage, 16 GB of RAM and at least 40 Tops (trillions of trillions per second).
Microsoft says that optimized Deepseek R1 models will allow developers to create artificial intelligence -supported applications and run them efficiently on the device. The company also states that with the use of low -bit processing systems, it has implemented technologies that will enable R1 models to work locally.
However, Microsoft does not only integrate for Copilot+ computers. It also includes the Deepseek R1 on the Azure AI Foundry platform. This platform currently includes OpenAI’s GPT-4, Mistral AI and Artificial Intelligence models such as Meta-Llama 3. The inclusion of Microsoft’s Deepseek model is considered a remarkable development. Because there are allegations that the company investigates whether Deepseek’s OpenAI technology without permission.
Deepseek R1 has aroused a great impact in the artificial intelligence world with the need for low calculation power. The need for less equipment compared to competing models was a source of concern for some chip manufacturers. Especially in Nvidia’s shares, there was a decrease after this development.
Microsoft says that the Deepseek R1 has successfully passed the security assessments and is subjected to meticulous tests. The company emphasizes that the model is undergoing a comprehensive test process with automatic assessments, security audits and risk reduction procedures.
The Chinese -based Deepseek has become a rapidly growing actor in the artificial intelligence market. The company claims that it has completed the R1 in less than two months and at a cost of lower than $ 6 million, using only 2,048 artificial intelligence accelerator in the process of developing the R1. In addition, the free presentation of R1 API under open source MIT license stands out as another remarkable element in the industry.
These developments lead to questioning the huge investments made by large technology companies in artificial intelligence models. Because Deepseek is pioneering a significant change in the sector, including competition with less cost and hardware.
Microsoft offers a great advantage with the integration of Deepseek R1
The integration of Microsoft’s Deepseek R1 into both Windows Copilot+ computers and Azure AI Foundry platform offers a great advantage for developers. Local employment of artificial intelligence -supported applications will make the user experience more efficient and will accelerate the spread of AI applications.
Microsoft’s support for Deepseek R1 shows that the competition in the field of artificial intelligence is further escalating. This model, which is offered with NPU optimization on Copilot+ computers, will allow AI -supported applications to work smoothly on the device. At the same time, the presence of the Azure AI Foundry platform will enable artificial intelligence to adopt in a wider ecosystem.