Back to leaderboard

qwen3.8-27b-mlx

LM-STUDIO MLX

qwen3_5 · 27B · 8bit

Marginal

Aug 17, 2026 · Apple M5 Max

llama3.2:3b

OLLAMA GGUF

llama · 3.2B · Q4_K_M

Good

Jun 25, 2026 · Cortex-A78AE

Global Score
59 vs 74
Hardware Fit
63 vs 99
Quality Score
58 vs 63

Hardware

qwen3.8-27b-mlx llama3.2:3b
MachineMacBook ProNVIDIA NVIDIA Jetson Orin Nano Engineering Reference Developer Kit Super
CPUApple M5 MaxCortex-A78AE
Cores186
RAM64 GB7 GB
GPUApple M5 MaxIntegrated / Unknown
OSmacOS 27.0Ubuntu 22.04.5 LTS
Archarm64arm64
Power Modebalancedbalanced

Performance

qwen3.8-27b-mlx llama3.2:3b
Tokens/sec18.321.1
First chunk4 ms1032 ms
TTFT530 ms1.0 s
Load timeN/A8.4 s
Memory usage38.5 GB2.4 GB
Memory %60%32%

HW Fit Score Breakdown

qwen3.8-27b-mlx

Speed
23/50
TTFT
20/20
Memory
20/30

llama3.2:3b

Speed
50/50
TTFT
20/20
Memory
29/30

Quality

qwen3.8-27b-mlx

Reasoning
19/20
Coding
14/20
Instruction
6/20
Structured
1/15
Math
13/15
Multilingual
5/10
Reasoning: Strong Coding: Adequate Instruction Following: Weak Structured Output: Poor Math: Strong Multilingual: Weak

llama3.2:3b

Reasoning
10/20
Coding
14/20
Instruction
15/20
Structured
12/15
Math
3/15
Multilingual
9/10
Reasoning: Adequate Coding: Adequate Instruction Following: Strong Structured Output: Strong Math: Poor Multilingual: Strong

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