Microsoft has revealed the newest addition to its Phi family of generative AI models.

Called Phi-4, the model improves in several areas over its predecessors, Microsoft claims, particularly in solving math problems. That’s partly the result of better training data quality.

Phi-4 is available in very limited access as of Thursday night only on Microsoft’s recently launched Azure AI Foundry development platform, and only for research purposes under a Microsoft research license agreement.

This is Microsoft’s latest small language model, coming in at 14 billion parameters in size, and it will compete with other small models such as GPT-4o mini, Gemini 2.0 Flash, and Claude 3.5 Haiku. These smaller AI models are often faster and cheaper to run, and their performance has gradually increased over the last several years.

In this case, Microsoft attributes Phi-4’s jump in performance to the use of “high-quality synthetic datasets” alongside high-quality datasets of human-generated content, and some unspecified post-training improvements.

Many AI labs are looking closely at the innovations they can make around synthetic data and post-training these days. Scale AI’s CEO Alexandr Wang said in a tweet on Thursday that “we have reached a pre-training data wall,” confirming several reports on the topic in the last several weeks.

Notably, Phi-4 is the first Phi-series model to launch following the departure of Sébastien Bubeck. Previously one of the vice presidents of AI at Microsoft and a key figure in the company’s Phi model development, Bubeck left the company in October to join OpenAI.