qwen3.5-4b-mlx

qwen3_5 · 4B · 4bit

MacBook Air (Apple M4)

32 GB · macOS 26.3

Tested on March 3, 2026 · Submitted by Topaz750

Top 42% Compare

Global Score

71 /100

Good

Hardware Fit

88/100

Quality

63/100

Get this model

Get it in LM Studio

Search and download models directly from the app

🤗

Find on HuggingFace

MLX versions & conversions

Hardware

Machine: MacBook Air
CPU: Apple M4
Cores: 10 total (4 perf + 6 eff)
Frequency: 2.4 GHz
RAM: 32 GB LPDDR5
GPU: Apple M4
OS: macOS 26.3
Arch: arm64
Power Mode: balanced

Performance

Tokens/sec Tokens generated per second — higher is better: 23.1
Standard deviation Variation of speed across benchmark runs — lower means more consistent: ±18.9
First chunk latency Delay before the first streamed chunk arrives from runtime: 3.3 s
Time to first token How fast the model starts responding — lower is better: 3.3 s
Load time Time to load the model into memory before inference: N/A
Memory usage RAM consumed during inference — percentage of total system memory: 2.9 GB (9%)
Total tokens Total number of tokens generated across all benchmark prompts: 832

Score breakdown

Speed

45/50

Time to first token

13/20

Memory

30/30

Quality

Reasoning

15/20

Coding

16/20

Instruction following

6/20

Structured output

6/15

Math

11/15

Multilingual

9/10

Category levels

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

Metadata

Spec version: 0.2.0
Runtime: LM Studio 0.4.6+1
Model format: MLX
Hardware profile: BALANCED
Result hash: ea30a4abf804ea75baa3144c550e24e77d3b2ae547a14b2f9576b803c25521fb

Interpretation

Hardware fit: 88/100. Overall suitability: GOOD (Global 71/100). Category profile: Reasoning: Adequate, Coding: Strong, Instruction Following: Weak, Structured Output: Weak, Math: Adequate, Multilingual: Strong.

Warnings

Token speed is unstable (stddev 18.9 tok/s, mean 23.1 tok/s) — may indicate thermal throttling or memory pressure.

Bench Environment

Power: AC CPU load: avg 12% (peak 19%)

Run yours now

$ npm install -g metrillm@latest

$ metrillm

Requires Node 20+ and Ollama or LM Studio running

Or run without installing: npx metrillm@latest