gemma3:4b
OLLAMA GGUFgemma3 · 4.3B · Q4_K_M
Jun 25, 2026 · Cortex-A78AE
GPT-20B:20b
OLLAMA GGUFgptoss · 20.9B · MXFP4
Aug 22, 2026 · NVIDIA RTX Spark N1X (6144-core GPU, 20-core ) RTX Spark N1X (6144-core GPU, 20-core )
Global Score
46 vs 93
Hardware Fit
93 vs 91
Quality Score
26 vs 94
Hardware
gemma3:4b GPT-20B:20b
MachineNVIDIA NVIDIA Jetson Orin Nano Engineering Reference Developer Kit SuperHP HyperX OMEN Transcend Gaming Laptop 14-ga0xxx5CD612DXML
CPUCortex-A78AENVIDIA RTX Spark N1X (6144-core GPU, 20-core ) RTX Spark N1X (6144-core GPU, 20-core )
Cores620
RAM7 GB55 GB
GPUIntegrated / UnknownNVIDIA RTX Spark N1X (6144-core Blackwell RTX GPU), NVIDIA NPU
OSUbuntu 22.04.5 LTSMicrosoft Windows 11 Pro 10.0.28000
Archarm64arm64
Power Modebalancedbalanced
Performance
gemma3:4b GPT-20B:20b
Tokens/sec14.957.8
First chunk2204 ms203 ms
TTFT2.2 s3.0 s
Load time2.0 s19.2 s
Memory usage2.6 GB12.0 GB
Memory %35%22%
HW Fit Score Breakdown
gemma3:4b
Speed
45/50
TTFT
19/20
Memory
29/30
GPT-20B:20b
Speed
50/50
TTFT
11/20
Memory
30/30
Quality
gemma3:4b
Reasoning
11/20
Coding
11/20
Instruction
0/20
Structured
0/15
Math
4/15
Multilingual
0/10
Reasoning: Adequate Coding: Adequate Instruction Following: Poor Structured Output: Poor Math: Weak Multilingual: Poor
GPT-20B:20b
Reasoning
19/20
Coding
17/20
Instruction
18/20
Structured
15/15
Math
15/15
Multilingual
10/10
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