MS
MICROSOFT / Phi-4-multimodal-instruct

Phi‑4 Multimodal

Compact Microsoft model that combines language, vision, and audio, including speech transcription and translation.

01 / SUMMARY

The essential

ORIENTATIONCompact multimodal

Primary use stated or inferred cautiously from official documentation.

CONTEXT128,000 tokens

Maximum input capacity when published by the source.

EXITNot published

Documented maximum generation limit.

ACCESSWeights and providers

Verified availability channels.

MODEL TYPEText and reasoning · Vision and documents · Voice and audio

Primary functional family and verified specialties.

LICENSEMIT

Declared terms for API access, model weights, or self-hosting.

02 / TECHNICAL SHEET

Limits and integration

API IDPhi-4-multimodal-instruct
Model typeText and reasoning · Vision and documents · Voice and audio
Access modelOpen weights
LicenseMIT
DeploymentLocal, edge, or private cloud
ReleaseFEB 2025
Knowledge cutoffNot published
EntranceText · Image · Audio
ExitText
Context window128,000 tokens
maximum outputNot published
ReasoningNot applicable or not published
Published tools
Structured OutingsNot applicable or not published
Batch processingNot published
Prompt cacheNot published
Fine-tuningDownloadable weights
Verified platformsHugging Face · Microsoft Foundry
BEST SUITED FOR

Compact multimodal applications, edge, vision, and voice.

WORTH MONITORING

Validate actual memory requirements, remote repository code, and performance by modality.

03 / CAPABILITIES

What it can do

01

Orientation

Compact multimodal applications, edge, vision, and voice.

02

Context

128,000 tokens · Not published

03

Tools and integration

Not published

04

Access

Weights and providers

Input modalities
TextImageAudio
04 / PRICES

Documented cost

ConceptWorthUnit/condition
Standard inputNot publishedStandard API rate
Cached inputNot publishedReading reused prefixes
Cache write or storageNot publishedThe condition varies by provider
Standard outputNot publishedMay include reasoning tokens
Batch inputNot publishedAsynchronous processing
Batch outputNot publishedAsynchronous processing

Consult the primary source: the billing unit depends on the model type and access channel.

i Prices change and may depend on level, region or context length. Check the source before making a decision.

05 / EVALUATIONS

How to read the results

“

A public benchmark provides guidance, but does not replace an evaluation with your data, tools, budget, and error tolerance.

BenchmarkResultMetricSource
Evaluaciones Phi‑4 MultimodalPublishedLanguage, vision, and voiceView source ↗

Inferama only highlights a “best result” when the metric, test set, configuration, and date allow for an equivalent comparison. The supplier's figures are presented as claims from its own source.

06 / VERSIONS

Chronology

Phi‑4 Multimodal

Version added to Inferama's verified catalog.

07 / SOURCES

Traceability