- Developers seeking a compact, MIT-licensed reasoning model for local deployment
- Teams that need top MMLU performance in the sub-20B parameter range
- Builders looking for a lightweight open model that runs efficiently on consumer hardware
Models
Phi-4
Microsoft's compact 14B dense reasoning model, MIT-licensed, tops MMLU in its size class with 16K context.
12K Stars
MIT License
Open sourceLocal firstSelf-hosted
About
Phi-4 overview
Microsoft's compact 14B dense reasoning model, MIT-licensed, tops MMLU in its size class with 16K context.
Compare
How it compares
Evaluate Phi-4 against nearby options by workload fit, license, and deployment model.
FAQ
Questions
What should I check before using Phi-4?
Run your own prompt set across reasoning, coding, latency, context length, and license constraints.
Is Phi-4 open source?
Phi-4 is listed on OpenAgent.bot with MIT based on the current resource metadata. Re-check the official repository, docs, and license before production use.
Tags
Capabilities
Decision brief
JSON
Should you use Phi-4?
- Use cases requiring very long context beyond 16K tokens
- Workloads that need a large MoE architecture with hundreds of billions of parameters
- Verified 2026-06-16
- License: MIT
- No GitHub repo recorded
- Open-source signal
Deployment
local, self hosted, cloud
Permission surface
memory
Decision signals
Local first, Self-hostable
Agent packet
Structured decision data for Phi-4
This packet is the compact machine-readable view agents should use before following source links or taking action.
Capabilities
local inference, inference
Constraints
open source, open weights, self hosted, local first
Deployment
local, self hosted, cloud
Permission surface
memory
Recommended workflows
Local or private AI stack
Overview
What Phi-4 does
What it is
Phi-4 is a model in the models category. Microsoft's compact 14B dense reasoning model, MIT-licensed, tops MMLU in its size class with 16K context.
Why it matters
Phi-4 delivers state-of-the-art reasoning performance in the 14B parameter class with a permissive MIT license, making it the go-to compact model for developers who need strong reasoning on modest hardware.
How to evaluate it
Evaluate Phi-4 by starting from the official sources, checking its docs, demo interface surface, and running one narrow workflow before expanding scope. Recorded integrations include GitHub, Hugging Face.
Facts
Known metadata and operating surface
These fields are separated from editorial interpretation so agents can reason over facts and missing checks.
Resource type model
Category Models
Maturity active
Difficulty Unknown
License MIT
Pricing open source
Verified 2026-06-16
Source confidence low
Risk level low
Fit matrix
Where Phi-4 fits in an agent stack
strong
Local or private AI stack
Phi-4 has multiple signals for local or private ai stack, including matching tags, capabilities, category, or positioning.
- Verify hardware requirements, data path, storage, and whether all calls stay in your environment.
- Confirm official docs, current maintenance, license, and runtime constraints before production use.
partial
Browser automation
Phi-4 has at least one signal for browser automation, but should be checked against a real task before adoption.
- Run one non-sensitive website task and inspect clicks, waits, retries, and changed URLs.
- Confirm official docs, current maintenance, license, and runtime constraints before production use.
partial
Memory or RAG workflow
Phi-4 has at least one signal for memory or rag workflow, but should be checked against a real task before adoption.
- Create, update, retrieve, correct, and delete memory or retrieval objects with real data.
- Confirm official docs, current maintenance, license, and runtime constraints before production use.
weak
Coding agent workflow
Phi-4 is not primarily positioned for coding agent workflow in the current metadata.
- Run a small repository change and inspect the diff, tests, and rollback path.
- Confirm official docs, current maintenance, license, and runtime constraints before production use.
weak
Connector or protocol layer
Phi-4 is not primarily positioned for connector or protocol layer in the current metadata.
- Connect one low-risk service, then inspect schemas, auth scope, errors, and logs.
- Confirm official docs, current maintenance, license, and runtime constraints before production use.
weak
Evaluation and observability
Phi-4 is not primarily positioned for evaluation and observability in the current metadata.
- Add one repeatable test case and confirm results can run again in review or CI.
- Confirm official docs, current maintenance, license, and runtime constraints before production use.
Inputs and outputs
What an agent should inspect
Likely inputs
- Prompts, messages, documents, images, or model inputs
- Official setup instructions and a small real workflow
Likely outputs
- A decision on whether this resource fits the target workflow
Evidence
Sources, claims, and missing checks
Claims are marked separately from source links so future crawlers and reviewers can update them without rewriting the page.
Phi-4 is listed as open source.
License metadata: MITPhi-4 supports these recorded deployment modes: local, self hosted, cloud.
OpenAgent decision signal metadata.Phi-4 is tagged with local inference, inference capabilities.
OpenAgent capability taxonomy.- GitHub repository has not been recorded.
- Dedicated docs link is missing.
- Repository freshness has not been recorded.
Next action
How to start evaluating Phi-4
Open Hugging Face
Start from the official source before adopting third-party instructions.
Open source Compare
Alternatives and nearby resources
Use related resources to compare category fit, license, deployment model, and first-workflow behavior.
FAQ
Common questions about Phi-4
What is Phi-4 used for?
Phi-4 is used as a model for models workflows. The most relevant recorded capabilities are local inference, inference.
Is Phi-4 open source?
Phi-4 is listed as open source with MIT license metadata. Re-check the official repository or source link before production use.
Can agents use Phi-4 directly?
Phi-4 has recorded interfaces such as docs, demo. Agents should prefer the JSON or Markdown profile first, then follow official docs for real execution.
What should I check before production use?
Check source confidence (low), risk level (low), license, maintenance freshness, permission surface, required credentials, and whether the first workflow succeeds in a sandbox.