qwen3.8-27b-mlx
LM-STUDIO MLXqwen3_5 · 27B · 4bit
Oct 20, 2026 · Apple M1 Pro
codestral:22b
OLLAMA GGUFllama · 22.2B · Q4_0
Jun 26, 2026 · Cortex-X925
Global Score
78 vs 67
Hardware Fit
56 vs 67
Quality Score
87 vs 67
Hardware
qwen3.8-27b-mlx codestral:22b
MachineMacBook ProNVIDIA NVIDIA_DGX_Spark
CPUApple M1 ProCortex-X925
Cores1020
RAM32 GB122 GB
GPUApple M1 ProDevice 2e12
OSmacOS 15.6.1Ubuntu 24.04.4 LTS
Archarm64arm64
Power Modebalancedperformance
Performance
qwen3.8-27b-mlx codestral:22b
Tokens/sec10.817.2
First chunk8 ms129 ms
TTFT2.6 s129 ms
Load timeN/A11.6 s
Memory usage21.0 GB18.9 GB
Memory %66%16%
HW Fit Score Breakdown
qwen3.8-27b-mlx
Speed
23/50
TTFT
16/20
Memory
17/30
codestral:22b
Speed
17/50
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
codestral:22b
Reasoning
11/20
Coding
17/20
Instruction
9/20
Structured
14/15
Math
6/15
Multilingual
10/10
Reasoning: Adequate Coding: Strong Instruction Following: Weak Structured Output: Strong Math: Weak 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