qwen3.5:9b-mlx
OLLAMA GGUFnvfp4
Aug 26, 2026 · Apple M1 Pro
Kimi-48B:48B
OLLAMA GGUFkimi-linear · 49.1B · unknown
Aug 19, 2026 · Intel Core™ Ultra X7 358H
Global Score
54 vs 86
Hardware Fit
81 vs 78
Quality Score
43 vs 89
Hardware
qwen3.5:9b-mlx Kimi-48B:48B
MachineMacBook ProGMKtec NucBox EVO-T2S
CPUApple M1 ProIntel Core™ Ultra X7 358H
Cores1016
RAM16 GB64 GB
GPUApple M1 ProIntel(R) Arc(TM) B390 GPU
OSmacOS 26.6.2Microsoft Windows 11 Pro 10.0.26200
Archarm64x64
Power Modebalancedbalanced
Performance
qwen3.5:9b-mlx Kimi-48B:48B
Tokens/sec25.222.8
First chunk341 ms721 ms
TTFT6.3 s721 ms
Load time3.4 s23.0 s
Memory usage8.4 GB25.5 GB
Memory %52%40%
HW Fit Score Breakdown
qwen3.5:9b-mlx
Speed
50/50
TTFT
7/20
Memory
24/30
Kimi-48B:48B
Speed
31/50
TTFT
20/20
Memory
27/30
Quality
qwen3.5:9b-mlx
Reasoning
14/20
Coding
2/20
Instruction
4/20
Structured
4/15
Math
10/15
Multilingual
9/10
Reasoning: Adequate Coding: Poor Instruction Following: Poor Structured Output: Weak Math: Adequate Multilingual: Strong
Kimi-48B:48B
Reasoning
18/20
Coding
19/20
Instruction
15/20
Structured
15/15
Math
13/15
Multilingual
9/10
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