qwen3.5:latest

qwen35 · 9.7B · Q4_K_M

THINKING MODEL

Mac mini (Apple M4)

16 GB · macOS 26.3

Tested on March 8, 2026

Top 95% Compare

Global Score

34 /100

Not Rec.

Hardware Fit

58/100

Quality

23/100

Get this model

🦙

Ollama

ollama pull qwen3.5:latest

View on Ollama Library

ollama.com/library/qwen3.5

Get it in LM Studio

Search and download models directly from the app

🤗

Find on HuggingFace

GGUF versions & quantizations

Hardware

Machine: Mac mini
CPU: Apple M4
Cores: 10 total (4 perf + 6 eff)
Frequency: 2.4 GHz
RAM: 16 GB LPDDR5
GPU: Apple M4
OS: macOS 26.3
Arch: arm64
Power Mode: balanced

Performance

Tokens/sec Tokens generated per second — higher is better: 13.3
Standard deviation Variation of speed across benchmark runs — lower means more consistent: ±0.0
First chunk latency Delay before the first streamed chunk arrives from runtime: 438 ms
Time to first token How fast the model starts responding — lower is better: 19.8 s
Load time Time to load the model into memory before inference: 0.2 s
Memory usage RAM consumed during inference — percentage of total system memory: 8.0 GB (50%)
Total tokens Total number of tokens generated across all benchmark prompts: 1429
Thinking tokens (est.) Estimated internal reasoning tokens — models with "thinking" generate more tokens: ~727

Score breakdown

Speed

32/50

Time to first token

1/20

Memory

25/30

Quality

Reasoning

6/20

Coding

1/20

Instruction following

3/20

Structured output

0/15

Math

5/15

Multilingual

8/10

Category levels

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

Metadata

Spec version: 0.2.1
Runtime: Ollama 0.17.4
Model format: GGUF
Hardware profile: ENTRY
Result hash: 329f59ac7ff00d3cc70ffecbf467ec8d277e8894162d0dfa89306cfa82125e48

Interpretation

Hardware fit: 58/100. Overall suitability: NOT RECOMMENDED (Global 34/100). Category profile: Reasoning: Weak, Coding: Poor, Instruction Following: Poor, Structured Output: Poor, Math: Weak, Multilingual: Strong.

Warnings

Significant swap activity during benchmark (+0.6 GB). Model may exceed available RAM — results are severely degraded.

Bench Environment

Power: AC Swap delta: +0.6 GB CPU load: avg 8% (peak 11%)

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