Microsoft: Phi 4

Public pricingIntelligence 76/100Small memoryTool use

Microsoft: Phi 4 is a text model for reasoning and problem solving. It combines deep reasoning and planning and low latency and efficient inference with a 16K tokens context window and a low-cost profile. Use it for reasoning, analysis, and hard problem solving when latency, cost, and throughput matters.

Input

$0.07/1M

Output

$0.14/1M

Cached

$0.01/1M

Batch

$0.07/1M

Calculate your Phi 4 bill.

Set your workload — see cost at your exact volume.

What would Phi 4 cost you?

Adjust the workload to see your monthly bill.

1,00010,00050,000250,0001M10M

Technical specifications

Phi 4 at a glance.

Memory

16,384

tokens

Max reply

16,384

tokens

Memory tier

Small

a few emails or a short document

Tokenizer

—

Released

Dec 2024

Training cutoff

Jun 2024

Availability

Public pricing

Status

active

Benchmarks

Quality benchmarks

Independent evaluations from public leaderboards. Higher is better.

  • mmlu

    84.8

  • gpqa_diamond

    56.1

  • humaneval

    82.6

  • math

    80.4

  • aa_intelligence_index

    10

  • ifeval

    4.88

  • bbh

    52.58

  • mmlu_pro

    47.72

What it can do

Capabilities & limits.

  • Understands images
  • Deep step-by-step thinking
  • Uses tools / calls functions
  • Strict JSON output
  • Streams replies
  • Fine-tunable on your data

When to pick Phi 4

  • Agentic workflows that call tools or APIs.
  • High-volume workloads where unit cost matters.

When to look elsewhere

  • Your workload involves images — pick a vision-capable model instead.
  • Your inputs routinely exceed short documents.

Alternatives

Similar models to compare.

FAQ

Phi 4 — the questions we see most.

Pricing, capabilities, alternatives — generated from the same data that powers the calculator above.

Get instant answers from our AI agent

At a typical workload of 50,000 conversations a month with 1,500-token prompts and 800-token replies, Phi 4 costs roughly $10 per month. Input is $0.07 /1M tokens and output is $0.14 /1M tokens.
Phi 4 has a 16,384-token context window (small memory — a few emails or a short document). That means you can fit about 3,072 words of input and history in a single call.
Beyond text generation, Phi 4 supports calling functions / tools, strict JSON output, fine-tuning on your own data. It streams replies by default.
Phi 4 was released in December 2024, with training data cut off around June 2024.
Models in a similar class include WizardLM-2 8x22B, Qwen3.5-Flash, Nova Lite 1.0. The "Similar models" section below this FAQ links into each.

Still unsure?

Compare Phi 4 against 100+ other models.

Open the full wizard — pick a use case, set your usage, and see side-by-side monthly costs in under a minute.

Open the comparison wizard