Back to leaderboard

microsoft/phi-4-mini-reasoning

LM-STUDIO MLX

phi3 · 3.8B · 4bit

Good

Mar 5, 2026 · Apple M4 Pro

qwen2.5-math-1.5b-instruct

LM-STUDIO MLX

qwen2 · 1.5B · 4bit

Marginal

Mar 4, 2026 · Apple M4

Global Score
64 vs 45
Hardware Fit
100 vs 99
Quality Score
48 vs 22

Hardware

microsoft/phi-4-mi… qwen2.5-math-1.5b-…
MachineMac miniMacBook Air
CPUApple M4 ProApple M4
Cores1410
RAM64 GB32 GB
GPUApple M4 ProApple M4
OSmacOS 15.7.4macOS 26.3
Archarm64arm64
Power Modebalancedbalanced

Performance

microsoft/phi-4-mi… qwen2.5-math-1.5b-…
Tokens/sec90.391.7
First chunk198 ms205 ms
TTFT198 ms205 ms
Load timeN/AN/A
Memory usage2.0 GB0.8 GB
Memory %3%3%

HW Fit Score Breakdown

microsoft/phi-4-mini-re…

Speed
50/50
TTFT
20/20
Memory
30/30

qwen2.5-math-1.5b-instr…

Speed
50/50
TTFT
20/20
Memory
29/30

Quality

microsoft/phi-4-mini-re…

Reasoning
7/20
Coding
7/20
Instruction
7/20
Structured
7/15
Math
13/15
Multilingual
7/10
Reasoning: Weak Coding: Weak Instruction Following: Weak Structured Output: Weak Math: Strong Multilingual: Adequate

qwen2.5-math-1.5b-instr…

Reasoning
2/20
Coding
0/20
Instruction
3/20
Structured
1/15
Math
14/15
Multilingual
2/10
Reasoning: Poor Coding: Poor Instruction Following: Poor Structured Output: Poor Math: Strong Multilingual: Poor

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