nvidia/nemotron-3-nano
LM-STUDIO MLXnemotron_h · 30B · 4bit
Mar 4, 2026 · Apple M4
qwen3.5-0.8b-mlx
LM-STUDIO MLXqwen3_5 · 0.8B · 4bit
Mar 6, 2026 · Apple M4 Pro
Global Score
84 vs 53
Hardware Fit
100 vs 89
Quality Score
77 vs 38
Hardware
nvidia/nemotron-3-… qwen3.5-0.8b-mlx
MachineMacBook AirMac mini
CPUApple M4Apple M4 Pro
Cores1014
RAM32 GB64 GB
GPUApple M4Apple M4 Pro
OSmacOS 26.3macOS 15.7.4
Archarm64arm64
Power Modebalancedbalanced
Performance
nvidia/nemotron-3-… qwen3.5-0.8b-mlx
Tokens/sec47.952.9
First chunk465 ms4013 ms
TTFT465 ms4.0 s
Load timeN/AN/A
Memory usage16.6 GB0.6 GB
Memory %52%1%
HW Fit Score Breakdown
nvidia/nemotron-3-nano
Speed
50/50
TTFT
20/20
Memory
30/30
qwen3.5-0.8b-mlx
Speed
50/50
TTFT
9/20
Memory
30/30
Quality
nvidia/nemotron-3-nano
Reasoning
13/20
Coding
18/20
Instruction
14/20
Structured
15/15
Math
7/15
Multilingual
10/10
Reasoning: Adequate Coding: Strong Instruction Following: Adequate Structured Output: Strong Math: Weak Multilingual: Strong
qwen3.5-0.8b-mlx
Reasoning
7/20
Coding
2/20
Instruction
12/20
Structured
10/15
Math
2/15
Multilingual
5/10
Reasoning: Weak Coding: Poor Instruction Following: Adequate Structured Output: Adequate Math: Poor Multilingual: Adequate
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