Back to leaderboard

qwen3.5-2b-mlx

LM-STUDIO MLX

qwen3_5 · 2B · 4bit

Good

Mar 6, 2026 · Apple M4 Pro

gemma4:12b-mlx

OLLAMA GGUF
Excellent

Jul 4, 2026 · Apple M2 Pro

Global Score
65 vs 84
Hardware Fit
88 vs 82
Quality Score
55 vs 85

Hardware

qwen3.5-2b-mlx gemma4:12b-mlx
MachineMac miniMacBook Pro
CPUApple M4 ProApple M2 Pro
Cores1410
RAM64 GB16 GB
GPUApple M4 ProApple M2 Pro
OSmacOS 15.7.4macOS 26.5
Archarm64arm64
Power Modebalancedbalanced

Performance

qwen3.5-2b-mlx gemma4:12b-mlx
Tokens/sec48.927.2
First chunk4208 ms732 ms
TTFT4.2 s8.1 s
Load timeN/A3.5 s
Memory usage1.6 GB7.1 GB
Memory %3%44%

HW Fit Score Breakdown

qwen3.5-2b-mlx

Speed
50/50
TTFT
8/20
Memory
30/30

gemma4:12b-mlx

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

Quality

qwen3.5-2b-mlx

Reasoning
10/20
Coding
7/20
Instruction
13/20
Structured
14/15
Math
3/15
Multilingual
8/10
Reasoning: Adequate Coding: Weak Instruction Following: Adequate Structured Output: Strong Math: Poor Multilingual: Strong

gemma4:12b-mlx

Reasoning
16/20
Coding
17/20
Instruction
13/20
Structured
15/15
Math
14/15
Multilingual
10/10
Reasoning: Strong Coding: Strong Instruction Following: Adequate Structured Output: Strong Math: Strong 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