gemma4:31b
OLLAMA GGUFgemma4 · 30.7B · Q4_K_M
Oct 2, 2026 · AMD Ryzen 9 5950X 16-Core Processor
qwopus3.6-35b-a3b-coder-mtp
LM-STUDIO GGUFqwen35moe · 35B · Q4_K_M
Jul 4, 2026 · AMD Ryzen 7 7800X3D 8-Core Processor
Global Score
64 vs 84
Hardware Fit
41 vs 79
Quality Score
74 vs 86
Hardware
gemma4:31b qwopus3.6-35b-a3b-…
MachineSystem manufacturerASUS
CPUAMD Ryzen 9 5950X 16-Core ProcessorAMD Ryzen 7 7800X3D 8-Core Processor
Cores3216
RAM32 GB31 GB
GPUNVIDIA GeForce RTX 3090NVIDIA GeForce RTX 4090, Raphael
OSMicrosoft Windows 11 Pro 10.0.26300cachyos rolling
Archx64x64
Power Modebalancedlow-power
Performance
gemma4:31b qwopus3.6-35b-a3b-…
Tokens/sec9.5106.2
First chunk1751 ms3 ms
TTFT30.0 s271 ms
Load time33.9 sN/A
Memory usage1.4 GB25.0 GB
Memory %4%82%
HW Fit Score Breakdown
gemma4:31b
Speed
11/50
TTFT
0/20
Memory
30/30
qwopus3.6-35b-a3b-coder…
Speed
50/50
TTFT
20/20
Memory
9/30
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
qwopus3.6-35b-a3b-coder…
Reasoning
16/20
Coding
19/20
Instruction
14/20
Structured
15/15
Math
12/15
Multilingual
10/10
Reasoning: Strong Coding: Strong Instruction Following: Adequate 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