gemma4:latest
OLLAMA GGUFgemma4 · 8.0B · Q4_K_M
Jun 5, 2026 · AMD Ryzen 7 5800H
qwen3:4b
OLLAMA GGUFqwen3 · 4.0B · Q4_K_M
Jun 26, 2026 · Cortex-A78AE
Global Score
89 vs 29
Hardware Fit
83 vs 86
Quality Score
92 vs 4
Hardware
gemma4:latest qwen3:4b
MachineLENOVO 82JQNVIDIA NVIDIA Jetson Orin Nano Engineering Reference Developer Kit Super
CPUAMD Ryzen 7 5800HCortex-A78AE
Cores166
RAM16 GB7 GB
GPUNVIDIA GeForce RTX 3060 Laptop GPUIntegrated / Unknown
OSMicrosoft Windows 11 家庭版 中文版 10.0.26100Ubuntu 22.04.5 LTS
Archx64arm64
Power Modebalancedbalanced
Performance
gemma4:latest qwen3:4b
Tokens/sec46.317.9
First chunk824 ms1010 ms
TTFT3.8 s6.4 s
Load time21.0 s9.5 s
Memory usage8.9 GB3.0 GB
Memory %56%40%
HW Fit Score Breakdown
gemma4:latest
Speed
50/50
TTFT
11/20
Memory
22/30
qwen3:4b
Speed
50/50
TTFT
8/20
Memory
28/30
Quality
gemma4:latest
Reasoning
18/20
Coding
19/20
Instruction
15/20
Structured
15/15
Math
15/15
Multilingual
10/10
Reasoning: Strong Coding: Strong Instruction Following: Strong Structured Output: Strong Math: Strong Multilingual: Strong
qwen3:4b
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
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