microsoft/phi-4-mini-reasoning

LM-STUDIO MLX

phi3 · 3.8B · 4bit

Good

Mar 3, 2026 · Apple M4

qwen3-4b-qwen3.6-plus-reasoning-distilled

LM-STUDIO GGUF

qwen3 · 4B · Q4_1

Good

Apr 21, 2026 · Apple M2 Max

Global Score

62 vs 65

Hardware Fit

99 vs 96

Quality Score

46 vs 51

Hardware

microsoft/phi-4-mi… qwen3-4b-qwen3.6-p…

MachineMacBook AirMac Studio

CPUApple M4Apple M2 Max

Cores1012

RAM32 GB96 GB

GPUApple M4Apple M2 Max

OSmacOS 26.3macOS 26.1

Archarm64arm64

Power Modebalancedbalanced

Performance

microsoft/phi-4-mi… qwen3-4b-qwen3.6-p…

Tokens/sec40.399.6

First chunk289 ms18 ms

TTFT289 ms1.6 s

Load timeN/A4.8 s

Memory usage2.0 GB0.3 GB

Memory %6%0%

HW Fit Score Breakdown

microsoft/phi-4-mini-re…

Speed

50/50

TTFT

20/20

Memory

29/30

qwen3-4b-qwen3.6-plus-r…

Speed

50/50

TTFT

16/20

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

qwen3-4b-qwen3.6-plus-r…

Reasoning

8/20

Coding

4/20

Instruction

14/20

Structured

9/15

Math

8/15

Multilingual

8/10

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

View microsoft/phi-4-mi… details | View qwen3-4b-qwen3.6-p… 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