Back to leaderboard

qwen3.5-4b-mlx

LM-STUDIO MLX

qwen3_5 · 4B · 4bit

Good

Mar 3, 2026 · Apple M4

qwen2.5:72b

OLLAMA GGUF

qwen2 · 72.7B · Q4_K_M

Not Rec.

Jun 27, 2026 · Cortex-X925

Global Score
71 vs 75
Hardware Fit
88 vs 50
Quality Score
63 vs 86

Hardware

qwen3.5-4b-mlx qwen2.5:72b
MachineMacBook AirNVIDIA NVIDIA_DGX_Spark
CPUApple M4Cortex-X925
Cores1020
RAM32 GB122 GB
GPUApple M4Device 2e12
OSmacOS 26.3Ubuntu 24.04.4 LTS
Archarm64arm64
Power Modebalancedperformance

Performance

qwen3.5-4b-mlx qwen2.5:72b
Tokens/sec23.14.4
First chunk3279 ms529 ms
TTFT3.3 s529 ms
Load timeN/A49.7 s
Memory usage2.9 GB54.0 GB
Memory %9%44%

HW Fit Score Breakdown

qwen3.5-4b-mlx

Speed
45/50
TTFT
13/20
Memory
30/30

qwen2.5:72b

Speed
4/50
TTFT
20/20
Memory
26/30

Quality

qwen3.5-4b-mlx

Reasoning
15/20
Coding
16/20
Instruction
6/20
Structured
6/15
Math
11/15
Multilingual
9/10
Reasoning: Adequate Coding: Strong Instruction Following: Weak Structured Output: Weak Math: Adequate Multilingual: Strong

qwen2.5:72b

Reasoning
19/20
Coding
19/20
Instruction
12/20
Structured
14/15
Math
12/15
Multilingual
10/10
Reasoning: Strong Coding: Strong Instruction Following: Adequate Structured Output: Strong Math: Strong Multilingual: Strong
View qwen3.5-4b-mlx details | View qwen2.5:72b 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