Microsoft
Reasoning

Phi-4

Frontier SLM

Microsoft Research's 14B reasoning small model. Punches well above its weight on math and code, and runs on a single GPU.

Phi-4 is a reasoning AI model from Microsoft. It costs $0.0700 per million input tokens and $0.1400 per million output tokens (blended $0.0910/M), with a 16,000-token context window.

RatesUSD per million tokens. Blend is 70% input, 30% output, our calculation.

INPUT
$0.0700/M
per million input tokens
OUTPUT
$0.1400/M
per million output tokens
BLENDED 70/30
$0.0910/M

0.0%over 91 days · hover to read

Phi-4 — blended price

Reconstructed from 91 days of recorded rate-card changes. Prices hold flat between changes because that is what a posted price does — no value here is interpolated.

CONTEXT
16,000
tokens

From OpenRouter's listing, read on 17 Nov 2026.

What it is good at
  • 14B fits one GPU
  • Strong math/code for size
  • Permissive license
Typical use cases
  • On-prem reasoning
  • Edge inference
  • Distillation studies

Benchmarks

vs. best public score
MMLU84.8%
Multitask academic knowledge across 57 subjects. Source: Microsoft (self-reported), MMLU, read 2026-11-10.
Graduate-level science questions, "Google-proof". Source: Microsoft (self-reported), GPQA Diamond, read 2026-11-10.
MATH80%
High-school competition math problems. No source recorded yet.
HumanEval82.6%
Python function synthesis from docstrings. Source: Microsoft (self-reported), HumanEval, read 2026-11-10.
Hand-curated from each provider's published reports and public leaderboards. Methodology varies across sources, treat as directional rather than authoritative.

More from Microsoft

See all 8 →

Price history

1 recorded change
Recorded price changes for Phi-4, newest first
DateRateWasNowChange
24 Jun 2026Input$0.0650$0.0700+7.7%

Rates per million tokens, as published by the provider or its router listing. Tracking runs since 3 May; anything earlier is not on record. See every tracked price change.

Frequently asked questions

How much does Phi-4 cost?

Phi-4 costs $0.0700 per million input tokens and $0.1400 per million output tokens, for a blended reference rate of $0.0910 per million tokens.

What is Phi-4's context window?

Phi-4 supports up to 16,000 tokens of context in a single request.

What is Phi-4 best for?

Phi-4 is well suited to 14B fits one GPU, Strong math/code for size and Permissive license.

Who makes Phi-4?

Phi-4 is developed and served by Microsoft.

Compared with

Terms used on this page