qwen/qwen3.8-27b
LM-STUDIO MLXqwen3_5 · 27B · 4bit
Aug 19, 2026 · Apple M4 Max
gemma4:12b
OLLAMA GGUFgemma4 · 11.9B · Q4_K_M
Jul 7, 2026 · Cortex-X925
Global Score
83 vs 86
Hardware Fit
68 vs 79
Quality Score
90 vs 89
Hardware
qwen/qwen3.8-27b gemma4:12b
MachineMacBook ProASUSTeK COMPUTER INC. GX10
CPUApple M4 MaxCortex-X925
Cores1620
RAM128 GB122 GB
GPUApple M4 MaxDevice 2e12
OSmacOS 26.5.2Ubuntu 24.04.4 LTS
Archarm64arm64
Power Modebalancedperformance
Performance
qwen/qwen3.8-27b gemma4:12b
Tokens/sec27.627.5
First chunk10 ms484 ms
TTFT4.6 s484 ms
Load timeN/A8.2 s
Memory usage21.0 GB9.3 GB
Memory %16%8%
HW Fit Score Breakdown
qwen/qwen3.8-27b
Speed
32/50
TTFT
6/20
Memory
30/30
gemma4:12b
Speed
29/50
TTFT
20/20
Memory
30/30
Quality
qwen/qwen3.8-27b
Reasoning
19/20
Coding
15/20
Instruction
16/20
Structured
15/15
Math
15/15
Multilingual
10/10
Reasoning: Strong Coding: Adequate Instruction Following: Strong Structured Output: Strong Math: Strong Multilingual: Strong
gemma4:12b
Reasoning
17/20
Coding
19/20
Instruction
17/20
Structured
15/15
Math
11/15
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