qwen3.5-9b-mlx

qwen3_5 · 9B · 4bit

THINKING MODEL

Mac mini (Apple M4)

24 GB · macOS 26.3

Tested on March 11, 2026 · Submitted by cryptoepops

Top 67% Compare

Global Score

59 /100

Marginal

Hardware Fit

65/100

Quality

57/100

Get this model

Get it in LM Studio

Search and download models directly from the app

🤗

Find on HuggingFace

MLX versions & conversions

Hardware

Machine: Mac mini
CPU: Apple M4
Cores: 10 total (4 perf + 6 eff)
Frequency: 2.4 GHz
RAM: 24 GB LPDDR5
GPU: Apple M4
OS: macOS 26.3
Arch: arm64
Power Mode: balanced

Performance

Tokens/sec Tokens generated per second — higher is better: 14.3
Standard deviation Variation of speed across benchmark runs — lower means more consistent: ±0.2
First chunk latency Delay before the first streamed chunk arrives from runtime: 24 ms
Time to first token How fast the model starts responding — lower is better: 13.4 s
Load time Time to load the model into memory before inference: 13.3 s
Memory usage RAM consumed during inference — percentage of total system memory: 2.5 GB (10%)
Total tokens Total number of tokens generated across all benchmark prompts: 1617
Thinking tokens (est.) Estimated internal reasoning tokens — models with "thinking" generate more tokens: ~614

Score breakdown

Speed

32/50

Time to first token

3/20

Memory

30/30

Quality

Reasoning

10/20

Coding

17/20

Instruction following

5/20

Structured output

9/15

Math

11/15

Multilingual

5/10

Category levels

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

Metadata

Spec version: 0.2.1
Runtime: LM Studio 0.4.6+1
Model format: MLX
Hardware profile: ENTRY
Result hash: f0e821ad6d4b50ee5fac6fe21f62611e4fa4b97cf081273c9274fd12882ff5dd

Interpretation

Hardware fit: 65/100. Overall suitability: MARGINAL (Global 59/100). Category profile: Reasoning: Weak, Coding: Strong, Instruction Following: Weak, Structured Output: Adequate, Math: Adequate, Multilingual: Adequate.

Warnings

Token throughput is estimated from LM Studio output because native token stats were unavailable. Compare tok/s across backends cautiously.

Bench Environment

Power: AC CPU load: avg 19% (peak 23%)

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