gemma4:12b-mlx

THINKING MODEL

MacBook Pro (Apple M2 Pro)

16 GB · macOS 26.5

Tested on July 4, 2026 · Submitted by enigmatracer

Top 25% Compare

Global Score

81 /100

Excellent

Hardware Fit

82/100

Quality

81/100

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 Pro
CPU: Apple M2 Pro
Cores: 10 total (6 perf + 4 eff)
Frequency: 2.4 GHz
RAM: 16 GB LPDDR5
GPU: Apple M2 Pro
OS: macOS 26.5
Arch: arm64
Power Mode: balanced

Performance

Tokens/sec Tokens generated per second — higher is better: 25.3
Standard deviation Variation of speed across benchmark runs — lower means more consistent: ±1.8
First chunk latency Delay before the first streamed chunk arrives from runtime: 761 ms
Time to first token How fast the model starts responding — lower is better: 7.7 s
Load time Time to load the model into memory before inference: 3.6 s
Memory usage RAM consumed during inference — percentage of total system memory: 7.1 GB (44%)
Total tokens Total number of tokens generated across all benchmark prompts: 1460
Thinking tokens (est.) Estimated internal reasoning tokens — models with "thinking" generate more tokens: ~704

Score breakdown

Speed

50/50

Time to first token

6/20

Memory

26/30

Quality

Reasoning

16/20

Coding

15/20

Instruction following

12/20

Structured output

15/15

Math

13/15

Multilingual

10/10

Category levels

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

Metadata

Spec version: 0.2.1
Runtime: Ollama 0.31.1
Model format: GGUF
Hardware profile: ENTRY
Result hash: de4f3b30c6df6956034f1d8a7263935528c57d2ee8d11cb721ab9a4dc7cba341

Interpretation

Hardware fit: 82/100. Overall suitability: EXCELLENT (Global 81/100). Category profile: Reasoning: Strong, Coding: Strong, Instruction Following: Adequate, Structured Output: Strong, Math: Strong, Multilingual: Strong.

Warnings

Significant swap activity during benchmark (+2.0 GB). Model may exceed available RAM — results are severely degraded.

Bench Environment

Power: AC Swap delta: +2.0 GB CPU load: avg 25% (peak 38%)

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