qwen3.8-27b-mlx
LM-STUDIO MLXqwen3_5 · 27B · 4bit
Oct 20, 2026 · Apple M1 Pro
cogito:3b
OLLAMA GGUFllama · 3.6B · Q4_K_M
Mar 6, 2026 · Apple M4
Global Score
78 vs 67
Hardware Fit
56 vs 100
Quality Score
87 vs 53
Hardware
qwen3.8-27b-mlx cogito:3b
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 cogito:3b
Tokens/sec10.845.6
First chunk8 ms177 ms
TTFT2.6 s177 ms
Load timeN/A0.8 s
Memory usage21.0 GB7.0 GB
Memory %66%22%
HW Fit Score Breakdown
qwen3.8-27b-mlx
Speed
23/50
TTFT
16/20
Memory
17/30
cogito:3b
Speed
50/50
TTFT
20/20
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
cogito:3b
Reasoning
6/20
Coding
8/20
Instruction
13/20
Structured
15/15
Math
2/15
Multilingual
9/10
Reasoning: Weak Coding: Weak Instruction Following: Adequate Structured Output: Strong Math: Poor 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