qwen2.5:3b

qwen2 · 3.1B · Q4_K_M

Good

Mar 1, 2026 · Apple M4

granite3.1-dense:2b

granite · 2.5B · Q4_K_M

Good

Mar 1, 2026 · Apple M4

Global Score

71 vs 67

Hardware Fit

99 vs 99

Quality Score

53 vs 45

Hardware

qwen2.5:3b granite3.1-dense:2b

MachineMacBook AirMacBook Air

CPUApple M4Apple M4

Cores1010

RAM32 GB32 GB

GPUApple M4Apple M4

OSmacOS 26.3macOS 26.3

Archarm64arm64

Power Modebalancedbalanced

Performance

qwen2.5:3b granite3.1-dense:2b

Tokens/sec38.553.4

TTFT204 ms122 ms

Load time0.5 s0.4 s

Memory usage4.0 GB5.6 GB

Memory %13%18%

HW Fit Score Breakdown

qwen2.5:3b

Speed

40/40

TTFT

30/30

Memory

29/30

granite3.1-dense:2b

Speed

40/40

TTFT

30/30

Memory

29/30

Quality

qwen2.5:3b

Reasoning

12/20

Coding

0/20

Instruction

12/20

Structured

15/15

Math

5/15

Multilingual

9/10

Reasoning: Adequate Coding: Poor Instruction Following: Adequate Structured Output: Strong Math: Weak Multilingual: Strong

granite3.1-dense:2b

Reasoning

8/20

Coding

0/20

Instruction

13/20

Structured

14/15

Math

3/15

Multilingual

7/10

Reasoning: Weak Coding: Poor Instruction Following: Adequate Structured Output: Strong Math: Poor Multilingual: Adequate

View qwen2.5:3b details | View granite3.1-dense:2b details

Run yours and compare

npx metrillm@latest bench

Requires Node 20+ and Ollama or LM Studio running