THE MODEL, IN DETAIL / Microsoft Research

phi-4

microsoft/phi-4Open weightsUnspecified source configuration configuration

[Microsoft Research](/microsoft) Phi-4 is designed to perform well in complex reasoning tasks and can operate efficiently in situations with limited memory or where quick responses are needed. At 14 billion...

OpenRouter

Source-linked model intelligenceCatalog added 10 Jan 2025Specifications observed 28 Oct 2026

Specifications and providers describe Microsoft: Phi 4. Evaluation scores retain the exact Unspecified source configuration configuration.

GPQA Diamond

56.1%

Source-reported

Context window

16.4K

14.7K max output tokens

Input price

$0.07

USD / million input tokens

Output price

$0.14

USD / million output tokens

Inference providers

2 provider routes observed

Measured output speed

—

No model-level measurement

01 / PERFORMANCE

A clearer view of capability.

Read the score, understand the context, and explore where this model sits among its peers.

Performance fingerprint

Individual published accuracy scores · not a composite index

Performance meets price

GPQA Diamond · same published source and version

Google DeepMindOpenAIAnthropicDeepSeekMicrosoft ResearchAlibaba

In good company

GPQA Diamond · Epoch AI · benchmark data

96.4

Google DeepMind

95.8

OpenAI

95.6

Anthropic

57.2

Google DeepMind

57.1

Google DeepMind,Google

56.8

Alibaba

56.5

DeepSeek

56.1

Alibaba

56.1

Mistral AI

56.1

This model

Rank #153 in this published cohort. 325 models in the source cohort. Showing selected nearby models. Evaluation settings can differ.

Evidence, by task.

Coverage is not a capability score.

3published results
1 evaluation source

Reasoning3 evaluations

CodingAwaiting evidence

Mathematics2 evaluations

VisionAwaiting evidence

AgenticAwaiting evidence

Long contextAwaiting evidence

Documented features, task evaluations, and preference rankings describe different things. Missing coverage does not imply a model cannot perform a task.

Explore our methodology

02 / BENCHMARK EXPLORER

The numbers behind the model.

Original results, exact versions, and a direct path to the evidence behind every score.

3 published results

3 percentage-based evaluations. Each benchmark measures a different task; compare results within the same evaluation and protocol.

Published benchmark results for phi-4
Version & metric	Source	Published
Epoch AI · independent runsAccuracy · Higher is better	Epoch AI · benchmark dataSource-reported	Not available
Epoch AI · independent runsAccuracy · Higher is better	Epoch AI · benchmark dataSource-reported	Not available
Epoch AI · independent runsAccuracy · Higher is better	Epoch AI · benchmark dataSource-reported	Not available

03 / COST & VALUE

Know what your workload costs.

Explore token prices, see where the spend goes, and model a workload of your own.

Input$0.07per million tokens

Output$0.14per million tokens

Cache read—per million tokens

Cache write—per million tokens

Workload cost calculator

Change the numbers. See the trade-offs.

Pricing source

Input tokens / requestOutput tokens / requestRequests

Cached share of input0%

Cache-read pricing is not documented for this selection.

Token estimate only. Excludes cache writes, tools, and other provider fees. Actual token usage and provider limits vary.

Your estimated spend

1,000 requests · 11M tokens

Uncached input$0.7

Cached input$0

Output$0.14

Per request$0.0008

Without cache discount$0.84

Estimated cache savings$0

04 / INFERENCE NETWORK

One model. Different ways to run it.

Compare inference routes by price, context, quantization, and reported availability.

1 providers · 2 routes

Standard service tier only

Provider routes for phi-4, observed 28 Oct 2026
Provider / route	Input / M	Output / M	Cache / M	Context	Output limit	Speed	Latency	Availability · 1d	Rates
DDeepInfraBF16 · flex	$0.056	$0.112	—	16.4K	14.7K	—	—	100%
DDeepInfraBF16	$0.07	$0.14	—	16.4K	14.7K	—	—	99.73%

Snapshot observed 28 Oct 2026. Speed and latency are the reported 30-minute median when available. This is a dated snapshot, not live monitoring.

Provider price comparison

USD per million tokens · input and output rates

InputOutput

DeepInfrabf16 · flex

$0.056

$0.112

DeepInfrabf16

$0.07

$0.14

Reported availability

Provider availability over the selected observation window

DeepInfra · flex · BF16100%

DeepInfra · BF1699.73%

Output speed

Reported 30-minute p50 · tokens per second

No provider throughput measurements were returned in this snapshot.

Response latency

Reported 30-minute p50 · seconds

No provider latency measurements were returned in this snapshot.

Context by provider

Actual route limits can differ from the catalog maximum

DeepInfra · flex · BF1616.4K

DeepInfra · BF1616.4K

05 / CAPABILITIES & MODALITIES

What goes in. What comes out.

A map of documented interfaces and API features, separate from benchmark performance.

INPUT MODALITIES

Text

phi-4Model interface

OUTPUT MODALITIES

Text

Feature support

Listed catalog API parameters and documented pricing features

ReasoningNot listed

Tool callingNot listed

Tool choiceNot listed

Structured outputsSupported

Response formatSupported

Reasoning detailsNot listed

Seed controlSupported

Stop sequencesSupported

Prompt cachingNot documented

TemperatureSupported

Room to work

Context capacity and completion limits

16.4Ktotal context tokens

Context window

14.7Kmax output

Output tokens share the total context budget. The maximum completion length is a separate limit, and individual provider routes may have smaller limits.

Catalog context: 16,384 tokens
Maximum output: 14,745 tokens
Top-provider context: 16,384 tokens

Features vary by route

Check the provider before depending on a specific API feature.

API feature support by provider
Provider route	Reasoning	Tools	Structured output	Response format	Implicit caching
DeepInfra · flex · BF16	Not listed	Not listed	Supported	Supported	Not listed
DeepInfra · BF16	Not listed	Not listed	Supported	Supported	Not listed

06 / UNDER THE HOOD

The complete specification sheet.

Exact identities, model architecture, limits, and the defaults that shape a request.

Identity & architecture

Developer: Microsoft Research
Exact identity: phi-4
Subject: Model Configuration
Access: Open weights
License: Not disclosed
Parameters: Not disclosed
Tokenizer: Other
Instruction format: Not disclosed

Reasoning & controls

Reasoning: Not listed
Mandatory reasoning: Not documented
Reasoning by default: Not documented
Published configuration: Unspecified source configuration
Supported reasoning efforts: Not documented
Default reasoning effort: Not disclosed
API parameters: 13
Provider moderation: No
Quantization variants: bf16

Dates & limits

Catalog added: 10 Jan 2025
Knowledge cutoff: 2024-06-30
Catalog expiration: Not documented
Specifications observed: 28 Oct 2026
Providers observed: 28 Oct 2026
Request limits: Provider-specific / not supplied
Canonical catalog ID: microsoft/phi-4

API parameters & defaults

Availability is catalog-wide. Defaults can vary across providers.

frequency_penaltySupported · default not supplied

logit_biasSupported · default not supplied

max_tokensSupported · default not supplied

min_pSupported · default not supplied

presence_penaltySupported · default not supplied

repetition_penaltySupported · default not supplied

response_formatSupported · default not supplied

seedSupported · default not supplied

stopSupported · default not supplied

structured_outputsSupported · default not supplied

temperatureSupported · default not supplied

top_kSupported · default not supplied

top_pSupported · default not supplied

07 / BUILD WITH THIS MODEL

From exploration to your first request.

Use the documented catalog identifier and keep your API key in an environment variable.

Make an API request

const response = await fetch(
  "https://openrouter.ai/api/v1/chat/completions",
  {
    method: "POST",
    headers: {
      Authorization: `Bearer ${process.env.OPENROUTER_API_KEY}`,
      "Content-Type": "application/json",
    },
    body: JSON.stringify({
      "model": "microsoft/phi-4",
      "messages": [
        {
          "role": "user",
          "content": "Explain the trade-offs in this design."
        }
      ]
    }),
  },
);
if (!response.ok) throw new Error(`HTTP ${response.status}`);
const data = await response.json();
console.log(data.choices[0].message.content);

08 / PROVENANCE & HISTORY

Every number has a story.

Follow the source, check the observation date, and distinguish published results from model specifications.

The source register

Direct links to the underlying records

OpenRouter model catalog

Architecture, modalities, parameters, and starting prices

Observed 28 Oct 2026

OpenRouter provider routes

Provider limits, prices, availability, and reported measurements

Observed 28 Oct 2026

Epoch AI · benchmark data

Source-reported · Epoch AI · independent runs

Observed 28 Oct 2026

The observation trail

Recorded events, in chronological order

10 Jan 2025Added to the model catalog

Catalog creation date, as recorded by OpenRouter.

28 Oct 2026Provider routes observed

2 inference routes in this snapshot.

28 Oct 2026Model specifications observed

Catalog pricing, limits, and supported parameters captured.

Observation dates describe when data was captured. They do not represent newly run evaluations or continuous performance monitoring.

Published observation ledger

Historical evidence, with source and version preserved.

Published observation history for phi-4
Observed	Benchmark	Version	Source	Verification
28 Oct 2026	OTIS Mock AIME 2024–2025	Epoch AI · independent runs	Epoch AI · benchmark data	Source-reported
28 Oct 2026	MATH Level 5	Epoch AI · independent runs	Epoch AI · benchmark data	Source-reported
28 Oct 2026	GPQA Diamond	Epoch AI · independent runs	Epoch AI · benchmark data	Source-reported
28 Oct 2026	OTIS Mock AIME 2024–2025	Epoch AI · independent runs	Epoch AI · benchmark data	Source-reported
28 Oct 2026	MATH Level 5	Epoch AI · independent runs	Epoch AI · benchmark data	Source-reported
28 Oct 2026	GPQA Diamond	Epoch AI · independent runs	Epoch AI · benchmark data	Source-reported
28 Oct 2026	OTIS Mock AIME 2024–2025	Epoch AI · independent runs	Epoch AI · benchmark data	Source-reported
28 Oct 2026	MATH Level 5	Epoch AI · independent runs	Epoch AI · benchmark data	Source-reported
28 Oct 2026	GPQA Diamond	Epoch AI · independent runs	Epoch AI · benchmark data	Source-reported

09 / A FEW MORE DETAILS

Questions worth asking.

Quick answers, drawn from this model’s connected specifications.

How much does phi-4 cost?

Catalog starting rates are $0.07 per million input tokens and $0.14 per million output tokens. Compare provider routes and service tiers above for their specific prices.

How much context can it handle?

16,384 tokens in the catalog context window. The catalog lists a maximum of 14,745 completion tokens. Provider limits can differ.

Does it support tools and structured output?

Tool calling: not listed. Structured outputs: listed as supported. See the provider feature matrix for route-specific differences.

How should I interpret these benchmark results?

Read the benchmark name, exact version, metric, source, and model configuration together. Different evaluation settings can change the result. Preference ratings and source-defined composite indices remain separate from task accuracy.

Are these prices and provider measurements live?

This page uses dated source snapshots. Catalog data was observed on 28 Oct 2026. Provider routes were observed on 28 Oct 2026. Missing measurements are not estimated or treated as zero.

Source-backed evidence. Different tasks. A clearer picture.

About Methodology Benchmark dashboard