qwen3:4b
OLLAMA GGUFqwen3 · 4.0B · Q4_K_M
Jun 26, 2026 · Cortex-A78AE
gemma4:31b
OLLAMA GGUFgemma4 · 31.3B · Q4_K_M
Apr 13, 2026 · Apple M4 Max
Global Score
29 vs 76
Hardware Fit
86 vs 46
Quality Score
4 vs 89
Hardware
qwen3:4b gemma4:31b
MachineNVIDIA NVIDIA Jetson Orin Nano Engineering Reference Developer Kit SuperMacBook Pro
CPUCortex-A78AEApple M4 Max
Cores616
RAM7 GB64 GB
GPUIntegrated / UnknownApple M4 Max
OSUbuntu 22.04.5 LTSmacOS 26.3.1
Archarm64arm64
Power Modebalancedbalanced
Performance
qwen3:4b gemma4:31b
Tokens/sec17.922.4
First chunk1010 ms565 ms
TTFT6.4 s30.0 s
Load time9.5 s6.9 s
Memory usage3.0 GB43.9 GB
Memory %40%69%
HW Fit Score Breakdown
qwen3:4b
Speed
50/50
TTFT
8/20
Memory
28/30
gemma4:31b
Speed
30/50
TTFT
0/20
Memory
16/30
Quality
qwen3:4b
Reasoning
4/20
Coding
0/20
Instruction
0/20
Structured
0/15
Math
0/15
Multilingual
0/10
Reasoning: Poor Coding: Poor Instruction Following: Poor Structured Output: Poor Math: Poor Multilingual: Poor
gemma4:31b
Reasoning
16/20
Coding
19/20
Instruction
16/20
Structured
15/15
Math
14/15
Multilingual
9/10
Reasoning: Strong Coding: Strong Instruction Following: Strong Structured Output: Strong Math: Strong 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