Phi Silica, the latest small language model announced by Microsoft

AI has been a focal point of this year’s Build conference, and we saw various AI-related innovations from Microsoft.

It seems that the company has no plans of stopping since they have announced a new language model.

Microsoft announces Phi Silica, a new small language model

Microsoft has its small language model called Phi-3, and the Phi-3 Vision model was recently announced as the latest member of the Phi family.

The company has also revealed another small language model called Phi Silica as Neowin reports. This new model will join the Windows Copilot Library, and unlike other models, it only has 3.3 billion parameters.

Here’s what a Microsoft spokesperson had to say regarding the new model:

It is optimized to run on the NPUs in Copilot+ PCs, bringing local inferencing and achieving first-token latency performance. Developers can access the Phi Silica API and deliver user experiences across the Windows ecosystem.​

Windows Copilot Library provides a rich set of functionalities that will enhance developers’ ability to take advantage of AI in Windows. Developers will be able to access Studio Effects, Live captions translations, Phi Silica, OCR and Recall User Activity APIs as part of the Windows App SDK release in June. More APIs like Text Summarization, Vector Embeddings and RAG API will be coming later.

As previously stated, Phi Silica will be a part of Copilot Library, and it has a set of APIs that is powered by more than 40 on-device models that will ship with Windows.

This is all part of the Copilot + PC brand, and if you want to get access to this feature, you’ll need a compatible device, such as one of the newly announced Snapdragon Surface laptops.