# 85 in LLM models

Phi-4

« A 14-billion-parameter language model that excels at complex reasoning, especially mathematics. Trained mainly on high-quality synthetic data, it outperforms larger models on certain specialized tasks »

Visit This Tool →