Skip to content

Phi 4 mini reasoning

microsoft/Phi-4-mini-reasoning
Open Source · chat · open-weights
Open Alert me on changes
Intelligence
Context
32.8K
Max output
Weights
Open
API $/1M
Modalities
text
Released
29 Apr 2025
Intelligence Index via Artificial Analysis · 0–100, higher is better
Download image Share on X Share on LinkedIn
AI summary
● machine-written

Phi-4-mini-reasoning: Microsoft's lightweight reasoning model

Phi-4-mini-reasoning is a lightweight open-weights model from Microsoft designed for reasoning tasks in math, science, and code. Built on synthetic data and filtered public sources with a 32.8K context window, it uses supervised fine-tuning and reinforcement learning for structured reasoning without requiring large parameter counts.

What's new
  • Trained with chain-of-thought traces and reinforcement learning targeting math, science, and code reasoning
  • Optimized for structured reasoning with high inference efficiency
  • Available as open-weights model under MIT license
Best for
Mathematical problem solvingScience reasoning tasksCode generation and analysis
Sources

Source: https://huggingface.co/microsoft/Phi-4-mini-reasoning