qwen3.8-27b-mlx
LM-STUDIO MLXqwen3_5 · 27B · 4bit
Oct 20, 2026 · Apple M1 Pro
granite3.1-dense:2b
OLLAMA GGUFgranite · 2.5B · Q4_K_M
Mar 2, 2026 · Apple M4
Global Score
78 vs 68
Hardware Fit
56 vs 100
Quality Score
87 vs 54
Hardware
qwen3.8-27b-mlx granite3.1-dense:2b
MachineMacBook ProMacBook Air
CPUApple M1 ProApple M4
Cores1010
RAM32 GB32 GB
GPUApple M1 ProApple M4
OSmacOS 15.6.1macOS 26.3
Archarm64arm64
Power Modebalancedbalanced
Performance
qwen3.8-27b-mlx granite3.1-dense:2b
Tokens/sec10.854.6
First chunk8 msN/A
TTFT2.6 s121 ms
Load timeN/A0.4 s
Memory usage21.0 GB5.6 GB
Memory %66%18%
HW Fit Score Breakdown
qwen3.8-27b-mlx
Speed
23/50
TTFT
16/20
Memory
17/30
granite3.1-dense:2b
Speed
40/50
TTFT
30/20
Memory
30/30
Quality
qwen3.8-27b-mlx
Reasoning
18/20
Coding
18/20
Instruction
15/20
Structured
15/15
Math
11/15
Multilingual
10/10
Reasoning: Strong Coding: Strong Instruction Following: Strong Structured Output: Strong Math: Adequate Multilingual: Strong
granite3.1-dense:2b
Reasoning
8/20
Coding
9/20
Instruction
13/20
Structured
14/15
Math
3/15
Multilingual
7/10
Reasoning: Weak Coding: Weak Instruction Following: Adequate Structured Output: Strong Math: Poor Multilingual: Adequate
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