Back to leaderboard

deepseek-coder-v2:16b

OLLAMA GGUF

deepseek2 · 15.7B · Q4_0

Excellent

Jun 26, 2026 · Cortex-X925

qwen3:4b

OLLAMA GGUF

qwen3 · 4.0B · Q4_K_M

Not Rec.

Jun 26, 2026 · Cortex-A78AE

Global Score
82 vs 29
Hardware Fit
100 vs 86
Quality Score
74 vs 4

Hardware

deepseek-coder-v2:… qwen3:4b
MachineNVIDIA NVIDIA_DGX_SparkNVIDIA NVIDIA Jetson Orin Nano Engineering Reference Developer Kit Super
CPUCortex-X925Cortex-A78AE
Cores206
RAM122 GB7 GB
GPUDevice 2e12Integrated / Unknown
OSUbuntu 24.04.4 LTSUbuntu 22.04.5 LTS
Archarm64arm64
Power Modeperformancebalanced

Performance

deepseek-coder-v2:… qwen3:4b
Tokens/sec117.217.9
First chunk275 ms1010 ms
TTFT275 ms6.4 s
Load time8.9 s9.5 s
Memory usage17.9 GB3.0 GB
Memory %15%40%

HW Fit Score Breakdown

deepseek-coder-v2:16b

Speed
50/50
TTFT
20/20
Memory
30/30

qwen3:4b

Speed
50/50
TTFT
8/20
Memory
28/30

Quality

deepseek-coder-v2:16b

Reasoning
8/20
Coding
19/20
Instruction
13/20
Structured
12/15
Math
14/15
Multilingual
8/10
Reasoning: Weak Coding: Strong Instruction Following: Adequate Structured Output: Strong Math: Strong Multilingual: Strong

qwen3:4b

Reasoning
4/20
Coding
0/20
Instruction
0/20
Structured
0/15
Math
0/15
Multilingual
0/10
Reasoning: Poor Coding: Poor Instruction Following: Poor Structured Output: Poor Math: Poor Multilingual: Poor

Run yours and compare

$ npm install -g metrillm@latest
$ metrillm

Requires Node 20+ and Ollama or LM Studio running

Or run without installing: npx metrillm@latest