impulse

Phi 4 Reasoning

Open weightsRetiring 17 Jan 2027
API id Phi-4-reasoning
Azure AI Foundry retires it on 17 Jan 2027 (in 60 days). No API serves it after that.
Input / 1M
$0.125
at Azure AI Foundry
Output / 1M
$0.50
at Azure AI Foundry
Context window
33K
tokens
Max output
4K
tokens per reply
Hosts
1
1 listings
Released
-
date not published

Pricing

List prices at Azure AI Foundry, in US dollars per million tokens.

TierInput / 1MOutput / 1M
Standard$0.125$0.50

What it costs a month

At Azure AI Foundry list prices, 30 days.

  • Chat assistant5,000 requests a day, 2K in / 0.5K out$75.00
  • RAG over documents2,000 requests a day, 8K in / 0.6K out$78.00
  • Coding agent1,000 requests a day, 40K in / 2K out$180
  • Overnight batch20,000 requests a day, 4K in / 0.4K out$420
Use your own numbers

Where to run it

1 host serve Phi 4 Reasoning. The comparison is on blended price, 3 parts input to 1 part output.

HostIn / out per 1M
Phi-4-reasoning
$0.125 / $0.50
Retires 17 Jan 2027

Capabilities

Core

  • Streaming
  • Function calling
  • JSON mode
  • Structured outputs

Multimodal

  • Vision
  • Image input
  • Audio
  • Image generation
  • Video

Advanced

  • Reasoning
  • Tool calling
  • MCP
  • Prompt caching
  • Batch API
  • Embeddings
Takes in
Text
Produces
Text
Open weights
License: MIT
Runtimes: Ollama, llama.cpp, vLLM

Cheaper alternatives

Current models that cost less and keep most of what this one can do: tools, vision, reasoning, structured output, caching and enough context.

Same model, cheaper host

No host sells it for less than Azure AI Foundry.

Smaller Microsoft models

No cheaper Microsoft model keeps its key capabilities.

Other frontier labs

  • Claude Haiku 5.5
    Anthropic · on Anthropic
    $0.10 / $0.50
    Has every key capability
    9% less
  • GPT-6 Luna
    OpenAI · on OpenAI
    $0.10 / $0.50
    Has every key capability
    9% less

Open-weight picks

  • Qwen3.8 Flash
    Alibaba Qwen · on AIHubMix
    $0.113 / $0.38
    Has every key capability
    18% less
  • Nemotron 3.5 Super VL
    NVIDIA · on Kyma API
    $0.081 / $0.472
    Has every key capability
    18% less
  • Gemma 4 31B IT
    Google · on W&B Inference
    $0.10 / $0.34
    Has every key capability
    27% less
  • DeepSeek V4.1 Flash
    DeepSeek · on Sail
    $0.09 / $0.30
    Has every key capability
    35% less

Call it

Using Azure AI Foundry's API.

Python
from litellm import completion

response = completion(
    model="azure_ai/Phi-4-reasoning",
    messages=[{"role": "user", "content": "Hello"}],
)
print(response.choices[0].message.content)

Quick answers

How much does Phi 4 Reasoning cost?

On Azure AI Foundry, Phi 4 Reasoning costs $0.125 per million input tokens and $0.50 per million output tokens.

Which APIs offer Phi 4 Reasoning?

Phi 4 Reasoning runs on 1 API: Azure AI Foundry.

What is Phi 4 Reasoning's context window?

33K tokens, with up to 4K tokens in one reply.

Is Phi 4 Reasoning being retired?

Azure AI Foundry retires it on 17 Jan 2027. No API serves it after that.

What are cheaper alternatives to Phi 4 Reasoning?

Models that cost less and keep most of what it can do: Claude Haiku 5.5 on Anthropic (9% less), GPT-6 Luna on OpenAI (9% less) and Qwen3.8 Flash on AIHubMix (18% less). Savings are on a mix of 3 input tokens to 1 output token.

What is Phi 4 Reasoning's API model id?

Use Phi-4-reasoning with Azure AI Foundry.

Sources

Raw catalog entry for Phi-4-reasoning on Azure AI Foundry
{
  "deprecation_date": "2027-01-17",
  "deprecation_source": "https://learn.microsoft.com/en-us/azure/foundry/concepts/model-lifecycle-retirement",
  "successor_model": "microsoft--phi-4",
  "input_cost_per_token": 1.25e-7,
  "litellm_provider": "azure_ai",
  "max_input_tokens": 32768,
  "max_output_tokens": 4096,
  "mode": "chat",
  "output_cost_per_token": 5e-7,
  "source": "https://prices.azure.com/api/retail/prices?$filter=serviceName%20eq%20'Foundry%20Models'%20and%20armRegionName%20eq%20'eastus'%20and%20priceType%20eq%20'Consumption'",
  "supports_function_calling": true,
  "supports_pdf_input": false,
  "supports_tool_choice": true,
  "supports_reasoning": true
}