Back to leaderboard

qwen/qwen3.8-27b

LM-STUDIO MLX

qwen3_5 · 27B · 4bit

Excellent

Aug 18, 2026 · Apple M4 Max

spark-x2.5-4b

LM-STUDIO GGUF

spark2_5 · 4B · Q4_K_M

Good

Sep 30, 2026 · AMD Ryzen 7 4800H

Global Score
87 vs 77
Hardware Fit
83 vs 90
Quality Score
88 vs 71

Hardware

qwen/qwen3.8-27b spark-x2.5-4b
MachineMacBook ProASUSTeK COMPUTER INC. ROG Strix G513IM_G513IM
CPUApple M4 MaxAMD Ryzen 7 4800H
Cores1616
RAM128 GB15 GB
GPUApple M4 MaxAMD Radeon(TM) Graphics, NVIDIA GeForce RTX 3060 Laptop GPU
OSmacOS 26.5.2Microsoft Windows 11 Pro 10.0.26200
Archarm64x64
Power Modebalancedbalanced

Performance

qwen/qwen3.8-27b spark-x2.5-4b
Tokens/sec29.148.9
First chunk9 ms20 ms
TTFT650 ms168 ms
Load timeN/AN/A
Memory usage21.0 GB9.4 GB
Memory %16%61%

HW Fit Score Breakdown

qwen/qwen3.8-27b

Speed
33/50
TTFT
20/20
Memory
30/30

spark-x2.5-4b

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

Quality

qwen/qwen3.8-27b

Reasoning
18/20
Coding
18/20
Instruction
16/20
Structured
15/15
Math
11/15
Multilingual
10/10
Reasoning: Strong Coding: Strong Instruction Following: Strong Structured Output: Strong Math: Adequate Multilingual: Strong

spark-x2.5-4b

Reasoning
13/20
Coding
14/20
Instruction
16/20
Structured
15/15
Math
4/15
Multilingual
9/10
Reasoning: Adequate Coding: Adequate Instruction Following: Strong Structured Output: Strong Math: Weak Multilingual: Strong
View qwen/qwen3.8-27b details | View spark-x2.5-4b 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