Microsoft
Updated May 2026
Phi
Small models that punch above their weight
Phi is Microsoft's small-language-model family: compact open-weight models trained on carefully selected data, tuned for reasoning, multimodal understanding, and on-device or low-cost deployment. Phi is not trying to be the largest model in the room. It is a reminder that data quality, narrow scope, and efficient architecture can matter as much as raw parameter count.
Variants
4
Best at
small open-weight deployment
Max context
128K
tokens
Pricing
Open weights
$0
per self-hosted model weights
Azure AI Foundry
varies
per hosted deployment
Variants
Sort the table by context window or cost to find the right variant. Click any version below for a battle card with ranks, pricing notes, and official links.
| Modalities | ||||
|---|---|---|---|---|
Phi-4 Mini Phi-4-mini-instruct | 128K | varies / varies | 2025 | textcode |
Phi-4 Multimodal Phi-4-multimodal-instruct | 128K | varies / varies | 2025 | textvisionaudio |
Phi-4 Reasoning microsoft/Phi-4-reasoning | 33K | varies / varies | 2025 | textcode |
Phi-4 Reasoning Vision 15B microsoft/Phi-4-reasoning-vision-15B | 16K | varies / varies | 2026 | textvision |
Battle card
Context rank
#3
within Phi
Capability rank
#1
modalities + reasoning
Weights
Open
self-hostable if licensed
Best fights to pick
- visual math tutoring
- document and chart reasoning
- screen understanding
- edge-friendly multimodal assistants
Rankings are Tendril directory ranks, computed from the model data shown here. Public benchmark leaderboards change often, so official docs and current benchmark pages should be checked before buying or deploying.
Learn
Lessons about this model
Structured lessons that cover Phi directly or put it in context alongside its rivals.
Phi: Family Map
Phi: Flagship Picker
Phi: Cost And Context
Phi: Reasoning Mode
Phi: Multimodal Workflows
Phi: Coding And Software Tasks
Phi: Agents And Tool Use
Phi: Local Or Private Deployment
Phi: Benchmarks And Reality Checks
Phi: Safety And Limits
Phi: Prompt Recipes
Phi: Industry Use Cases
Phi: Compare Against Rivals
Phi: Build A Mini Workflow
Phi: Capstone Selection Memo
Check yourself
Quizzes
Short, mixed-difficulty quiz sets on Phi and its model family.
Phi Model Family Quiz
5 questions
A short quiz on where Phi fits, which variants matter, and how to choose it responsibly.
Start quiz →Hands-on
Try these prompts
Ready-made prompts that show Phi at its best. Use them in your own AI workspace, then compare the output with what you learned in Tendril.
Phi same-task bakeoff
BuildersCompare Phi against a cheaper and stronger rival on one real task.
Use Phi for this task, then compare with one cheaper model and one stronger model: summarize a 10-page policy document into risks, decisions, owners, and open questions.
Phi failure-mode log
CreatorsTrack where Phi fails instead of only saving the best output.
Run this task three times. For every answer, mark failures as factual error, formatting error, weak reasoning, refusal, or missing citation.
Print & keep
Printable reference
One-page summaries and flowcharts — great for desks, classrooms, or study sessions.
Go deeper
Official resources
Straight from the lab — docs, API references, and the chat surfaces you can try today.
Strengths
- small open-weight deployment
- math and science reasoning
- edge and classroom-friendly footprint
Limits
- not a frontier generalist
- shorter context on reasoning-vision variant
- English-heavy training emphasis