Back to leaderboard

google/gemma-4-e4b

LM-STUDIO GGUF

gemma4 · Q4_K_M

Good

Jul 7, 2026 · AMD Ryzen 7 5800H

llama3.2:3b

OLLAMA GGUF

llama · 3.2B · Q4_K_M

Good

Jun 25, 2026 · Cortex-A78AE

Global Score
62 vs 74
Hardware Fit
77 vs 99
Quality Score
55 vs 63

Hardware

google/gemma-4-e4b llama3.2:3b
MachineLENOVO 82JQNVIDIA NVIDIA Jetson Orin Nano Engineering Reference Developer Kit Super
CPUAMD Ryzen 7 5800HCortex-A78AE
Cores166
RAM16 GB7 GB
GPUNVIDIA GeForce RTX 3060 Laptop GPUIntegrated / Unknown
OSMicrosoft Windows 11 家庭版 中文版 10.0.26100Ubuntu 22.04.5 LTS
Archx64arm64
Power Modebalancedbalanced

Performance

google/gemma-4-e4b llama3.2:3b
Tokens/sec27.621.1
First chunk23 ms1032 ms
TTFT480 ms1.0 s
Load timeN/A8.4 s
Memory usage13.2 GB2.4 GB
Memory %83%32%

HW Fit Score Breakdown

google/gemma-4-e4b

Speed
48/50
TTFT
20/20
Memory
9/30

llama3.2:3b

Speed
50/50
TTFT
20/20
Memory
29/30

Quality

google/gemma-4-e4b

Reasoning
18/20
Coding
17/20
Instruction
5/20
Structured
3/15
Math
11/15
Multilingual
1/10
Reasoning: Strong Coding: Strong Instruction Following: Weak Structured Output: Poor Math: Strong Multilingual: Poor

llama3.2:3b

Reasoning
10/20
Coding
14/20
Instruction
15/20
Structured
12/15
Math
3/15
Multilingual
9/10
Reasoning: Adequate Coding: Adequate Instruction Following: Strong Structured Output: Strong Math: Poor Multilingual: Strong
View google/gemma-4-e4b details | View llama3.2:3b details

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