Back to leaderboard

qwen/qwen3-8b

LM-STUDIO MLX

qwen3 · 8B · 4bit

Excellent

Mar 5, 2026 · Apple M4 Pro

qwen2.5:1.5b

OLLAMA GGUF

qwen2 · 1.5B · Q4_K_M

Good

Jun 17, 2026 · ZHAOXIN KaiXian KX-7000 KaiXian KX-7000

Global Score
89 vs 65
Hardware Fit
95 vs 87
Quality Score
87 vs 56

Hardware

qwen/qwen3-8b qwen2.5:1.5b
MachineMac miniShanghai Zhaoxin Semiconductor Co., Ltd. ZXE CRB
CPUApple M4 ProZHAOXIN KaiXian KX-7000 KaiXian KX-7000
Cores148
RAM64 GB30 GB
GPUApple M4 ProZX C-1190
OSmacOS 15.7.4Microsoft Windows 11 Pro 10.0.26100
Archarm64x64
Power Modebalancedbalanced

Performance

qwen/qwen3-8b qwen2.5:1.5b
Tokens/sec33.416.8
First chunk257 ms941 ms
TTFT257 ms941 ms
Load timeN/A0.2 s
Memory usage4.3 GB1.0 GB
Memory %7%3%

HW Fit Score Breakdown

qwen/qwen3-8b

Speed
45/50
TTFT
20/20
Memory
30/30

qwen2.5:1.5b

Speed
37/50
TTFT
20/20
Memory
30/30

Quality

qwen/qwen3-8b

Reasoning
19/20
Coding
14/20
Instruction
16/20
Structured
15/15
Math
13/15
Multilingual
10/10
Reasoning: Strong Coding: Adequate Instruction Following: Strong Structured Output: Strong Math: Strong Multilingual: Strong

qwen2.5:1.5b

Reasoning
8/20
Coding
10/20
Instruction
12/20
Structured
13/15
Math
5/15
Multilingual
8/10
Reasoning: Weak Coding: Adequate Instruction Following: Adequate Structured Output: Strong Math: Weak Multilingual: Strong
View qwen/qwen3-8b details | View qwen2.5:1.5b 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