microsoft/phi-4-mini-reasoning

LM-STUDIO MLX

phi3 · 3.8B · 4bit

Good

Mar 3, 2026 · Apple M4

qwen2.5-7b-instruct

LM-STUDIO MLX

qwen2 · 7B · 4bit

Excellent

Mar 3, 2026 · Apple M4

Global Score

62 vs 84

Hardware Fit

99 vs 97

Quality Score

46 vs 78

Hardware

microsoft/phi-4-mi… qwen2.5-7b-instruct

MachineMacBook AirMacBook Air

CPUApple M4Apple M4

Cores1010

RAM32 GB32 GB

GPUApple M4Apple M4

OSmacOS 26.3macOS 26.3

Archarm64arm64

Power Modebalancedbalanced

Performance

microsoft/phi-4-mi… qwen2.5-7b-instruct

Tokens/sec40.324.4

First chunk289 ms464 ms

TTFT289 ms464 ms

Load timeN/AN/A

Memory usage2.0 GB4.0 GB

Memory %6%13%

HW Fit Score Breakdown

microsoft/phi-4-mini-re…

Speed

50/50

TTFT

20/20

Memory

29/30

qwen2.5-7b-instruct

Speed

47/50

TTFT

20/20

Memory

30/30

Quality

microsoft/phi-4-mini-re…

Reasoning

7/20

Coding

6/20

Instruction

7/20

Structured

7/15

Math

12/15

Multilingual

7/10

Reasoning: Weak Coding: Weak Instruction Following: Weak Structured Output: Weak Math: Strong Multilingual: Adequate

qwen2.5-7b-instruct

Reasoning

14/20

Coding

17/20

Instruction

15/20

Structured

15/15

Math

8/15

Multilingual

9/10

Reasoning: Adequate Coding: Strong Instruction Following: Strong Structured Output: Strong Math: Adequate Multilingual: Strong

View microsoft/phi-4-mi… details | View qwen2.5-7b-instruct details

Run yours and compare

$ npm install -g metrillm@latest

$ metrillm

Requires Node 20+ and Ollama or LM Studio running

Or run without installing: npx metrillm@latest