qwen2.5:3b
OLLAMA GGUFqwen2 · 3.1B · Q4_K_M
Jun 26, 2026 · Cortex-A78AE
qwen3.5:4b
OLLAMA GGUFqwen35 · 4.7B · Q4_K_M
Mar 5, 2026 · AMD RYZEN AI MAX+ 395
Global Score
77 vs 65
Hardware Fit
100 vs 73
Quality Score
67 vs 62
Hardware
qwen2.5:3b qwen3.5:4b
MachineNVIDIA NVIDIA Jetson Orin Nano Engineering Reference Developer Kit SuperAZW GTR Pro
CPUCortex-A78AEAMD RYZEN AI MAX+ 395
Cores632
RAM7 GB125 GB
GPUIntegrated / UnknownAMD Radeon 8060S
OSUbuntu 22.04.5 LTSUbuntu 24.04.4 LTS
Archarm64x64
Power Modebalancedlow-power
Performance
qwen2.5:3b qwen3.5:4b
Tokens/sec21.445.7
First chunk1006 ms180 ms
TTFT1.0 s30.0 s
Load time6.4 s5.3 s
Memory usage2.0 GB16.2 GB
Memory %27%13%
HW Fit Score Breakdown
qwen2.5:3b
Speed
50/50
TTFT
20/20
Memory
30/30
qwen3.5:4b
Speed
43/50
TTFT
0/20
Memory
30/30
Quality
qwen2.5:3b
Reasoning
12/20
Coding
14/20
Instruction
11/20
Structured
15/15
Math
6/15
Multilingual
9/10
Reasoning: Adequate Coding: Adequate Instruction Following: Adequate Structured Output: Strong Math: Weak Multilingual: Strong
qwen3.5:4b
Reasoning
17/20
Coding
1/20
Instruction
13/20
Structured
8/15
Math
13/15
Multilingual
10/10
Reasoning: Strong Coding: Poor Instruction Following: Adequate Structured Output: Adequate 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