qwen3:30b-a3b
OLLAMA GGUFqwen3moe · 30.5B · Q4_K_M
Jun 25, 2026 · Cortex-X925
gemma4:31b
OLLAMA GGUFgemma4 · 31.3B · Q4_K_M
Jul 7, 2026 · Cortex-X925
Global Score
84 vs 66
Hardware Fit
93 vs 40
Quality Score
80 vs 77
Hardware
qwen3:30b-a3b gemma4:31b
MachineNVIDIA NVIDIA_DGX_SparkASUSTeK COMPUTER INC. GX10
CPUCortex-X925Cortex-X925
Cores2020
RAM122 GB122 GB
GPUDevice 2e12Device 2e12
OSUbuntu 24.04.4 LTSUbuntu 24.04.4 LTS
Archarm64arm64
Power Modeperformanceperformance
Performance
qwen3:30b-a3b gemma4:31b
Tokens/sec92.710.7
First chunk286 ms983 ms
TTFT1.5 s20.3 s
Load time15.3 s8.8 s
Memory usage20.3 GB21.4 GB
Memory %17%18%
HW Fit Score Breakdown
qwen3:30b-a3b
Speed
50/50
TTFT
13/20
Memory
30/30
gemma4:31b
Speed
10/50
TTFT
0/20
Memory
30/30
Quality
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
gemma4:31b
Reasoning
14/20
Coding
17/20
Instruction
13/20
Structured
14/15
Math
10/15
Multilingual
9/10
Reasoning: Adequate Coding: Strong Instruction Following: Adequate Structured Output: Strong Math: Adequate 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