Back to leaderboard

qwen3-30b-a3b-thinking-2507-claude-4.5-sonnet-high-reasoning-distill-mlx

LM-STUDIO MLX

qwen3_moe · 30B · MXFP4

Excellent

Apr 20, 2026 · Apple M2 Max

gemma4:12b-mlx

OLLAMA GGUF
Not Rec.

Aug 24, 2026 · Apple M1 Pro

Global Score
97 vs 71
Hardware Fit
95 vs 76
Quality Score
98 vs 69

Hardware

qwen3-30b-a3b-thin… gemma4:12b-mlx
MachineMac StudioMacBook Pro
CPUApple M2 MaxApple M1 Pro
Cores1210
RAM96 GB16 GB
GPUApple M2 MaxApple M1 Pro
OSmacOS 26.1macOS 26.6.2
Archarm64arm64
Power Modebalancedbalanced

Performance

qwen3-30b-a3b-thin… gemma4:12b-mlx
Tokens/sec79.625.0
First chunk11 ms787 ms
TTFT1.7 s30.0 s
Load time7.1 s4.0 s
Memory usage0.0 GB7.1 GB
Memory %0%45%

HW Fit Score Breakdown

qwen3-30b-a3b-thinking-…

Speed
50/50
TTFT
15/20
Memory
30/30

gemma4:12b-mlx

Speed
50/50
TTFT
0/20
Memory
26/30

Quality

qwen3-30b-a3b-thinking-…

Reasoning
19/20
Coding
19/20
Instruction
20/20
Structured
15/15
Math
15/15
Multilingual
10/10
Reasoning: Strong Coding: Strong Instruction Following: Strong Structured Output: Strong Math: Strong Multilingual: Strong

gemma4:12b-mlx

Reasoning
14/20
Coding
10/20
Instruction
9/20
Structured
15/15
Math
11/15
Multilingual
10/10
Reasoning: Adequate Coding: Adequate Instruction Following: Weak Structured Output: Strong Math: Adequate Multilingual: Strong
View qwen3-30b-a3b-thin… details | View gemma4:12b-mlx 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