Back to leaderboard

qwen3.5-4b-mlx

LM-STUDIO MLX

qwen3_5 · 4B · 4bit

Not Rec.

Mar 6, 2026 · Apple M4 Pro

hf.co/unsloth/Qwen3.8-27B-GGUF:UD-IQ1_M

OLLAMA GGUF

qwen35 · 27.3B · unknown

Good

Aug 21, 2026 · AMD RYZEN AI MAX+ 395

Global Score
62 vs 68
Hardware Fit
35 vs 69
Quality Score
73 vs 67

Hardware

qwen3.5-4b-mlx hf.co/unsloth/Qwen…
MachineMac miniAZW GTR Pro
CPUApple M4 ProAMD RYZEN AI MAX+ 395
Cores1432
RAM64 GB125 GB
GPUApple M4 ProRadeon 8060S
OSmacOS 15.7.4Ubuntu 24.04.4 LTS
Archarm64x64
Power Modebalancedlow-power

Performance

qwen3.5-4b-mlx hf.co/unsloth/Qwen…
Tokens/sec0.021.1
First chunk5548 ms571 ms
TTFT5.5 s571 ms
Load timeN/A4.9 s
Memory usage2.9 GB22.7 GB
Memory %5%18%

HW Fit Score Breakdown

qwen3.5-4b-mlx

Speed
0/50
TTFT
6/20
Memory
29/30

hf.co/unsloth/Qwen3.8-2…

Speed
20/50
TTFT
19/20
Memory
30/30

Quality

qwen3.5-4b-mlx

Reasoning
17/20
Coding
17/20
Instruction
9/20
Structured
8/15
Math
13/15
Multilingual
9/10
Reasoning: Strong Coding: Strong Instruction Following: Weak Structured Output: Adequate Math: Strong Multilingual: Strong

hf.co/unsloth/Qwen3.8-2…

Reasoning
12/20
Coding
10/20
Instruction
10/20
Structured
15/15
Math
10/15
Multilingual
10/10
Reasoning: Adequate Coding: Weak Instruction Following: Adequate Structured Output: Strong Math: Adequate Multilingual: Strong

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