microsoft

Phi 4 14B

Released 2024-12-12Context 16.384Kphi-4

PipelineScore

70.3FEEDER

Ranked #50 of 52 models · 4th percentileAn even spread: no standout, no liability (69 to 71 across all five categories). Best-fit profile: Local-first.

Category breakdown

Score per category, normalized 0–100 against the v1 anchor.

Code

68.5

Reason

69.4

Tool Use

70.2

RAG

69.4

Speed

70.8

Strengths

Speed70.8

Tool Use70.2

Reason69.4

Same model, different rigs

Every submission of Phi 4 14B on the 0–100 scale. The spread is the point: where it runs changes what you get.

0255075100

Best 72.3 on lab-macbook-pro-m3-pro-18gb · lowest 67.9 on m3-pro-36gb · spread 4.5 pts across 4 runs. Hover a dot for its rig.

Sample tasks

A taste of what the test pack measures. Full prompts are private and rotated daily.

CodeDifficulty 1code-fib-1

Fibonacci function

Write a Python `fib(n)` returning the nth Fibonacci number, O(n).

ReasonDifficulty 1reason-math-1

Train meeting time

Two trains, opposite directions, given speeds and start times — when do they meet?

RAGDifficulty 2rag-extract-1

Extract metrics to JSON

From the context, extract net sales, operating margin, and free cash flow as a JSON object. Numbers only.

Tool UseDifficulty 2tool-schema-1

OpenAPI param selection

Given an OpenAPI schema with limit/offset/sort, fill JSON for 'next 50, recent first.'

RAGDifficulty 2rag-grounding-1

Refuses to fabricate

Context lacks the answer — does the model fabricate or correctly say it can't?

Compare with

Phi 4 14B vs openai/gpt-oss-20b Phi 4 14B vs DeepSeek R1 671B-A37B Phi 4 14B vs Qwen 3 235B-A22B MoE Phi 4 14B vs DeepSeek V3 671B-A37B Phi 4 14B vs DeepSeek Coder V2 236B