Gemma4-26B:26b
OLLAMA GGUFgemma4 · 25.8B · Q4_K_M
Aug 22, 2026 · NVIDIA RTX Spark N1X (6144-core GPU, 20-core ) RTX Spark N1X (6144-core GPU, 20-core )
llama3.2:3b
OLLAMA GGUFllama · 3.2B · Q4_K_M
Jun 25, 2026 · Cortex-A78AE
Global Score
92 vs 74
Hardware Fit
100 vs 99
Quality Score
89 vs 63
Hardware
Gemma4-26B:26b llama3.2:3b
MachineHP HyperX OMEN Transcend Gaming Laptop 14-ga0xxx5CD612DXMLNVIDIA NVIDIA Jetson Orin Nano Engineering Reference Developer Kit Super
CPUNVIDIA RTX Spark N1X (6144-core GPU, 20-core ) RTX Spark N1X (6144-core GPU, 20-core )Cortex-A78AE
Cores206
RAM55 GB7 GB
GPUNVIDIA RTX Spark N1X (6144-core Blackwell RTX GPU), NVIDIA NPUIntegrated / Unknown
OSMicrosoft Windows 11 Pro 10.0.28000Ubuntu 22.04.5 LTS
Archarm64arm64
Power Modebalancedbalanced
Performance
Gemma4-26B:26b llama3.2:3b
Tokens/sec61.321.1
First chunk185 ms1032 ms
TTFT185 ms1.0 s
Load time22.9 s8.4 s
Memory usage16.4 GB2.4 GB
Memory %30%32%
HW Fit Score Breakdown
Gemma4-26B:26b
Speed
50/50
TTFT
20/20
Memory
30/30
llama3.2:3b
Speed
50/50
TTFT
20/20
Memory
29/30
Quality
Gemma4-26B:26b
Reasoning
16/20
Coding
19/20
Instruction
17/20
Structured
14/15
Math
13/15
Multilingual
10/10
Reasoning: Strong Coding: Strong Instruction Following: Strong Structured Output: Strong Math: Strong Multilingual: Strong
llama3.2:3b
Reasoning
10/20
Coding
14/20
Instruction
15/20
Structured
12/15
Math
3/15
Multilingual
9/10
Reasoning: Adequate Coding: Adequate Instruction Following: Strong Structured Output: Strong Math: Poor 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