losclouds
Model · Microsoft

Phi-4 Mini

Tiny but capable 3.8B Phi model for on-device and low-latency inference.

Pricing

Model Overview

Tracked token-pricing fields for this model family. Empty pricing fields stay hidden until the source publishes them.

Model pricing
Price fieldValue
Input$0.04 / 1M tokens
Output$0.13 / 1M tokens
SourcePhi-4 Mini pricing
VerifiedApr 5, 2026 (High)

Surface

Capabilities

Input and output modalities, enabled feature flags, strengths, and tradeoffs.

Model capabilities
AttributeValues
Modalities
text
to
text
Capabilities
batchSupportfunctionCallingstructuredOutputs
Strengths128K context in 3.8B, On-device friendly
TradeoffsLower reasoning ceiling than Phi-4

References

Official Links

Canonical launch, documentation, pricing, and release-note URLs.

Coverage

Benchmark Coverage

Reported benchmark families, versions, scores, sources, and notes.

Benchmark coverage
BenchmarkVersionScoreDateSourceNotes

Lifecycle

Release History

Lifecycle transitions and release timeline for this model family.

Release history
ReleaseAliasLifecycleRelease DateDeprecationShutdownSummary
Phi-4 Miniphi-4-miniActiveMar 1, 2025Current published model family snapshot.

Surfaces

Host Coverage

Provider-specific hosting, context, pricing notes, feature differences, and provider-status context.

Host coverage
HostTypeContextPricing NoteDifferences
Azure AI Foundryfirst-party128.0K$0.04/$0.13 per MTok.

Migration

Migration Guidance

Documented migration summary, successor families, and known breaking changes.

Migration guidance
TopicDetails
SummaryTiny on-device tier. Upgrade to Phi-4 for more demanding tasks.
Replacement modelsphi-4

Timeline

Change Events

Cataloged model-family updates and source references.

Change events
DateTypeTitleDescriptionSource
Mar 1, 2025family_addedPhi-4 Mini publishedInitial public model family launch.Phi-4 Mini release notes

From Microsoft

Other models from Microsoft