Microsoft Launches Phi-3-mini, An Efficient Small Language AI Model In Cost

JAKARTA - Microsoft announced on Tuesday April 22, a cost-efficient small language artificial intelligence model that can perform tasks such as content creation and making social media posts while using smaller amounts of data.

The company introduced an artificial intelligence model called Phi-3-mini, which can surpass its model performance twice its size in a wide range of testing that evaluates language, programming, and mathematics capabilities.

Smaller artificial intelligence models are designed to perform simpler tasks, thereby facilitating use by companies with limited resources, Microsoft said.

A business can use Phi-3 to summarize key points from long documents, extract relevant insights and industrial trends from market research reports, Microsoft added.

The Phi-3-mini will soon be available in Microsoft Azure cloud service artificial intelligence model catalogs, the Uphold Face machine learning model platform, and Ollama, a framework for running models on local engines, the company said.

Last week, Microsoft invested $1.5 billion in UAE-based artificial intelligence company G42. The company also previously partnered with French startup Mistral AI to make their model available through its Azure cloud computing platform.