Back to leaderboard

qwen3.5:4b-mlx

OLLAMA GGUF

nvfp4

Marginal

Aug 24, 2026 · Apple M2

qwen3.8:27b-bf16

OLLAMA GGUF

qwen35 · 27.3B · BF16

Excellent

Aug 22, 2026 · AMD EPYC 7713 64-Core Processor

Global Score
53 vs 85
Hardware Fit
82 vs 77
Quality Score
41 vs 88

Hardware

qwen3.5:4b-mlx qwen3.8:27b-bf16
MachineMac miniDell Inc. PowerEdge XE8545
CPUApple M2AMD EPYC 7713 64-Core Processor
Cores8128
RAM8 GB1007 GB
GPUApple M2GA100 [A100 SXM4 80GB], GA100 [A100 SXM4 80GB], Integrated Matrox G200eW3 Graphics Controller, GA100 [A100 SXM4 80GB], GA100 [A100 SXM4 80GB]
OSmacOS 26.5.2Rocky Linux 8.10
Archarm64x64
Power Modebalancedunknown

Performance

qwen3.5:4b-mlx qwen3.8:27b-bf16
Tokens/sec24.727.2
First chunk302 ms237 ms
TTFT7.4 s237 ms
Load time3.2 s43.8 s
Memory usage3.8 GB68.8 GB
Memory %47%7%

HW Fit Score Breakdown

qwen3.5:4b-mlx

Speed
50/50
TTFT
6/20
Memory
26/30

qwen3.8:27b-bf16

Speed
27/50
TTFT
20/20
Memory
30/30

Quality

qwen3.5:4b-mlx

Reasoning
13/20
Coding
2/20
Instruction
4/20
Structured
4/15
Math
10/15
Multilingual
8/10
Reasoning: Adequate Coding: Poor Instruction Following: Poor Structured Output: Weak Math: Adequate Multilingual: Strong

qwen3.8:27b-bf16

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
View qwen3.5:4b-mlx details | View qwen3.8:27b-bf16 details

Run yours and compare

$ npm install -g metrillm@latest
$ metrillm

Requires Node 20+ and Ollama or LM Studio running

Or run without installing: npx metrillm@latest