Phi 4

by Microsoft

Instruct
Phi-4 is a state-of-the-art open model trained on synthetic data, filtered public content, and curated academic material to support advanced reasoning. It uses a dense, decoder-only Transformer architecture with 14 billion parameters. The model has been fine-tuned and aligned through supervised training and preference optimization, improving its instruction-following ability and safety.
Provider
Context Size
Max Output
Latency
Speed
Cost

Data reflects historical performance over the past days.

API Usage

Seamlessly integrate our API into your project by following these simple steps:

  1. Generate your API key from your profile.
  2. Copy the example code and replace the placeholder with your API key or see our documentation.

You can choose from three automatic provider selection preferences:

  • speed – Prioritizes the provider with the fastest response time.
  • cost – Selects the most cost-efficient provider.
  • balanced – Offers an optimal mix of speed and cost.

API Usage