qwen3.8-27b-mlx
LM-STUDIO MLXqwen3_5 · 27B · 4bit
Oct 20, 2026 · Apple M1 Pro
gpt-oss:20b
OLLAMA GGUFgptoss · 20.9B · MXFP4
Jul 3, 2026 · AMD RYZEN AI MAX+ 395
Global Score
78 vs 94
Hardware Fit
56 vs 91
Quality Score
87 vs 95
Hardware
qwen3.8-27b-mlx gpt-oss:20b
MachineMacBook ProFramework Desktop (AMD Ryzen AI Max 300 Series)
CPUApple M1 ProAMD RYZEN AI MAX+ 395
Cores1032
RAM32 GB64 GB
GPUApple M1 ProAMD Radeon(TM) 8060S Graphics
OSmacOS 15.6.1Microsoft Windows 11 Pro for Workstations 10.0.26200
Archarm64x64
Power Modebalancedbalanced
Performance
qwen3.8-27b-mlx gpt-oss:20b
Tokens/sec10.848.0
First chunk8 ms357 ms
TTFT2.6 s2.3 s
Load timeN/A0.2 s
Memory usage21.0 GB16.4 GB
Memory %66%26%
HW Fit Score Breakdown
qwen3.8-27b-mlx
Speed
23/50
TTFT
16/20
Memory
17/30
gpt-oss:20b
Speed
50/50
TTFT
11/20
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
gpt-oss:20b
Reasoning
19/20
Coding
19/20
Instruction
17/20
Structured
15/15
Math
15/15
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