Back to leaderboard

google/gemma-4-e4b

LM-STUDIO GGUF

gemma4 · Q4_K_M

Good

Jul 7, 2026 · AMD Ryzen 7 5800H

qwen3.5-0.8b-mlx

LM-STUDIO MLX

qwen3_5 · 0.8B · 4bit

Marginal

Mar 6, 2026 · Apple M4 Pro

Global Score
62 vs 53
Hardware Fit
77 vs 89
Quality Score
55 vs 38

Hardware

google/gemma-4-e4b qwen3.5-0.8b-mlx
MachineLENOVO 82JQMac mini
CPUAMD Ryzen 7 5800HApple M4 Pro
Cores1614
RAM16 GB64 GB
GPUNVIDIA GeForce RTX 3060 Laptop GPUApple M4 Pro
OSMicrosoft Windows 11 家庭版 中文版 10.0.26100macOS 15.7.4
Archx64arm64
Power Modebalancedbalanced

Performance

google/gemma-4-e4b qwen3.5-0.8b-mlx
Tokens/sec27.652.9
First chunk23 ms4013 ms
TTFT480 ms4.0 s
Load timeN/AN/A
Memory usage13.2 GB0.6 GB
Memory %83%1%

HW Fit Score Breakdown

google/gemma-4-e4b

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

qwen3.5-0.8b-mlx

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

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

qwen3.5-0.8b-mlx

Reasoning
7/20
Coding
2/20
Instruction
12/20
Structured
10/15
Math
2/15
Multilingual
5/10
Reasoning: Weak Coding: Poor Instruction Following: Adequate Structured Output: Adequate Math: Poor Multilingual: Adequate
View google/gemma-4-e4b details | View qwen3.5-0.8b-mlx 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