qwen3:4b
OLLAMA GGUFqwen3 · 4.0B · Q4_K_M
Jun 26, 2026 · Cortex-A78AE
lmstudio-community/Phi-4-reasoning-plus-MLX-4bit
LM-STUDIO GGUFMar 5, 2026 · Apple M4
Global Score
29 vs 33
Hardware Fit
86 vs 74
Quality Score
4 vs 15
Hardware
qwen3:4b lmstudio-community…
MachineNVIDIA NVIDIA Jetson Orin Nano Engineering Reference Developer Kit SuperMacBook Air
CPUCortex-A78AEApple M4
Cores610
RAM7 GB32 GB
GPUIntegrated / UnknownApple M4
OSUbuntu 22.04.5 LTSmacOS 26.3
Archarm64arm64
Power Modebalancedbalanced
Performance
qwen3:4b lmstudio-community…
Tokens/sec17.911.5
First chunk1010 ms670 ms
TTFT6.4 s670 ms
Load time9.5 sN/A
Memory usage3.0 GB3.0 GB
Memory %40%9%
HW Fit Score Breakdown
qwen3:4b
Speed
50/50
TTFT
8/20
Memory
28/30
lmstudio-community/Phi-…
Speed
24/50
TTFT
20/20
Memory
30/30
Quality
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
lmstudio-community/Phi-…
Reasoning
8/20
Coding
1/20
Instruction
1/20
Structured
1/15
Math
1/15
Multilingual
3/10
Reasoning: Weak Coding: Poor Instruction Following: Poor Structured Output: Poor Math: Poor Multilingual: Weak
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