qwen3-30b-a3b-thinking-2507-claude-4.5-sonnet-high-reasoning-distill-mlx
LM-STUDIO MLXqwen3_moe · 30B · MXFP4
Apr 22, 2026 · Apple M2 Max
qwen3.5-9b
LM-STUDIO MLXqwen3_5 · 9B · 4bit
Mar 3, 2026 · Apple M4
Global Score
95 vs 81
Hardware Fit
94 vs 84
Quality Score
96 vs 79
Hardware
qwen3-30b-a3b-thin… qwen3.5-9b
MachineMac StudioMacBook Air
CPUApple M2 MaxApple M4
Cores1210
RAM96 GB32 GB
GPUApple M2 MaxApple M4
OSmacOS 26.1macOS 26.3
Archarm64arm64
Power Modebalancedbalanced
Performance
qwen3-30b-a3b-thin… qwen3.5-9b
Tokens/sec80.420.4
First chunk46 ms3034 ms
TTFT1.8 s3.0 s
Load timeN/AN/A
Memory usage21.2 GB5.6 GB
Memory %22%17%
HW Fit Score Breakdown
qwen3-30b-a3b-thinking-…
Speed
50/50
TTFT
14/20
Memory
30/30
qwen3.5-9b
Speed
41/50
TTFT
14/20
Memory
29/30
Quality
qwen3-30b-a3b-thinking-…
Reasoning
19/20
Coding
18/20
Instruction
19/20
Structured
15/15
Math
15/15
Multilingual
10/10
Reasoning: Strong Coding: Strong Instruction Following: Strong Structured Output: Strong Math: Strong Multilingual: Strong
qwen3.5-9b
Reasoning
14/20
Coding
15/20
Instruction
15/20
Structured
15/15
Math
10/15
Multilingual
10/10
Reasoning: Adequate Coding: Strong Instruction Following: Strong Structured Output: Strong Math: Adequate 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