Gemma4-26B:26b
OLLAMA GGUFgemma4 · 25.8B · Q4_K_M
Aug 19, 2026 · Intel Core™ Ultra X7 358H
qwen3:14b
OLLAMA GGUFqwen3 · 14.8B · Q4_K_M
Mar 11, 2026 · Apple M4
Global Score
87 vs 83
Hardware Fit
81 vs 73
Quality Score
89 vs 87
Hardware
Gemma4-26B:26b qwen3:14b
MachineGMKtec NucBox EVO-T2SMac mini
CPUIntel Core™ Ultra X7 358HApple M4
Cores1610
RAM64 GB24 GB
GPUIntel(R) Arc(TM) B390 GPUApple M4
OSMicrosoft Windows 11 Pro 10.0.26200macOS 26.3
Archx64arm64
Power Modebalancedbalanced
Performance
Gemma4-26B:26b qwen3:14b
Tokens/sec23.111.1
First chunk978 ms500 ms
TTFT978 ms500 ms
Load time14.7 s7.3 s
Memory usage16.7 GB9.4 GB
Memory %26%39%
HW Fit Score Breakdown
Gemma4-26B:26b
Speed
31/50
TTFT
20/20
Memory
30/30
qwen3:14b
Speed
25/50
TTFT
20/20
Memory
28/30
Quality
Gemma4-26B:26b
Reasoning
16/20
Coding
19/20
Instruction
17/20
Structured
14/15
Math
13/15
Multilingual
10/10
Reasoning: Strong Coding: Strong Instruction Following: Strong Structured Output: Strong Math: Strong Multilingual: Strong
qwen3:14b
Reasoning
17/20
Coding
18/20
Instruction
15/20
Structured
15/15
Math
12/15
Multilingual
10/10
Reasoning: Strong Coding: Strong Instruction Following: Strong Structured Output: Strong Math: Strong Multilingual: Strong
Run yours and compare
$
npm install -g metrillm@latest$
metrillmRequires Node 20+ and Ollama or LM Studio running
Or run without installing: npx metrillm@latest