gemma4:31b
OLLAMA GGUFgemma4 · 30.7B · Q4_K_M
Oct 2, 2026 · AMD Ryzen 9 5950X 16-Core Processor
qwen3:30b-a3b
OLLAMA GGUFqwen3moe · 30.5B · Q4_K_M
Jun 25, 2026 · Cortex-X925
Global Score
64 vs 84
Hardware Fit
41 vs 93
Quality Score
74 vs 80
Hardware
gemma4:31b qwen3:30b-a3b
MachineSystem manufacturerNVIDIA NVIDIA_DGX_Spark
CPUAMD Ryzen 9 5950X 16-Core ProcessorCortex-X925
Cores3220
RAM32 GB122 GB
GPUNVIDIA GeForce RTX 3090Device 2e12
OSMicrosoft Windows 11 Pro 10.0.26300Ubuntu 24.04.4 LTS
Archx64arm64
Power Modebalancedperformance
Performance
gemma4:31b qwen3:30b-a3b
Tokens/sec9.592.7
First chunk1751 ms286 ms
TTFT30.0 s1.5 s
Load time33.9 s15.3 s
Memory usage1.4 GB20.3 GB
Memory %4%17%
HW Fit Score Breakdown
gemma4:31b
Speed
11/50
TTFT
0/20
Memory
30/30
qwen3:30b-a3b
Speed
50/50
TTFT
13/20
Memory
30/30
Quality
gemma4:31b
Reasoning
11/20
Coding
17/20
Instruction
10/20
Structured
15/15
Math
12/15
Multilingual
9/10
Reasoning: Adequate Coding: Strong Instruction Following: Adequate Structured Output: Strong Math: Strong Multilingual: Strong
qwen3:30b-a3b
Reasoning
16/20
Coding
10/20
Instruction
18/20
Structured
13/15
Math
13/15
Multilingual
10/10
Reasoning: Strong Coding: Adequate 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