qwen/qwen3.8-27b
LM-STUDIO MLXqwen3_5 · 27B · 4bit
Aug 18, 2026 · Apple M4 Max
GPT-20B:20b
OLLAMA GGUFgptoss · 20.9B · MXFP4
Aug 22, 2026 · NVIDIA RTX Spark N1X (6144-core GPU, 20-core ) RTX Spark N1X (6144-core GPU, 20-core )
Global Score
87 vs 93
Hardware Fit
83 vs 91
Quality Score
88 vs 94
Hardware
qwen/qwen3.8-27b GPT-20B:20b
MachineMacBook ProHP HyperX OMEN Transcend Gaming Laptop 14-ga0xxx5CD612DXML
CPUApple M4 MaxNVIDIA RTX Spark N1X (6144-core GPU, 20-core ) RTX Spark N1X (6144-core GPU, 20-core )
Cores1620
RAM128 GB55 GB
GPUApple M4 MaxNVIDIA RTX Spark N1X (6144-core Blackwell RTX GPU), NVIDIA NPU
OSmacOS 26.5.2Microsoft Windows 11 Pro 10.0.28000
Archarm64arm64
Power Modebalancedbalanced
Performance
qwen/qwen3.8-27b GPT-20B:20b
Tokens/sec29.157.8
First chunk9 ms203 ms
TTFT650 ms3.0 s
Load timeN/A19.2 s
Memory usage21.0 GB12.0 GB
Memory %16%22%
HW Fit Score Breakdown
qwen/qwen3.8-27b
Speed
33/50
TTFT
20/20
Memory
30/30
GPT-20B:20b
Speed
50/50
TTFT
11/20
Memory
30/30
Quality
qwen/qwen3.8-27b
Reasoning
18/20
Coding
18/20
Instruction
16/20
Structured
15/15
Math
11/15
Multilingual
10/10
Reasoning: Strong Coding: Strong Instruction Following: Strong Structured Output: Strong Math: Adequate Multilingual: Strong
GPT-20B:20b
Reasoning
19/20
Coding
17/20
Instruction
18/20
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