Back to leaderboard

qwen3-4b-qwen3.6-plus-reasoning-distilled

LM-STUDIO GGUF

qwen3 · 4B · Q4_1

Good

Apr 21, 2026 · Apple M2 Max

qwen2.5-1.5b-instruct

LM-STUDIO GGUF

qwen2 · 1.5B · Q4_K_M

Good

Mar 6, 2026 · Intel Core™ i5-5300U

Global Score
65 vs 61
Hardware Fit
96 vs 69
Quality Score
51 vs 58

Hardware

qwen3-4b-qwen3.6-p… qwen2.5-1.5b-instr…
MachineMac StudioLENOVO 20BUS00700
CPUApple M2 MaxIntel Core™ i5-5300U
Cores124
RAM96 GB16 GB
GPUApple M2 MaxIntel(R) HD Graphics 5500
OSmacOS 26.1Microsoft Windows 10 Professionnel 10.0.19045
Archarm64x64
Power Modebalancedbalanced

Performance

qwen3-4b-qwen3.6-p… qwen2.5-1.5b-instr…
Tokens/sec99.66.6
First chunk18 ms2277 ms
TTFT1.6 s2.3 s
Load time4.8 sN/A
Memory usage0.3 GB0.9 GB
Memory %0%6%

HW Fit Score Breakdown

qwen3-4b-qwen3.6-plus-r…

Speed
50/50
TTFT
16/20
Memory
30/30

qwen2.5-1.5b-instruct

Speed
23/50
TTFT
19/20
Memory
27/30

Quality

qwen3-4b-qwen3.6-plus-r…

Reasoning
8/20
Coding
4/20
Instruction
14/20
Structured
9/15
Math
8/15
Multilingual
8/10
Reasoning: Weak Coding: Poor Instruction Following: Adequate Structured Output: Adequate Math: Adequate Multilingual: Strong

qwen2.5-1.5b-instruct

Reasoning
9/20
Coding
9/20
Instruction
13/20
Structured
14/15
Math
5/15
Multilingual
8/10
Reasoning: Weak Coding: Weak Instruction Following: Adequate Structured Output: Strong Math: Weak Multilingual: Strong

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