deepseek-r1:14b

qwen2 · 14.8B · Q4_K_M

Not Rec.

Mar 2, 2026 · Apple M4

phi:2.7b

phi2 · 3B · Q4_0

Marginal

Mar 2, 2026 · Apple M4

Global Score

22 vs 59

Hardware Fit

50 vs 100

Quality Score

10 vs 41

Hardware

deepseek-r1:14b phi:2.7b

MachineMacBook AirMacBook Air

CPUApple M4Apple M4

Cores1010

RAM32 GB32 GB

GPUApple M4Apple M4

OSmacOS 26.3macOS 26.3

Archarm64arm64

Power Modebalancedbalanced

Performance

deepseek-r1:14b phi:2.7b

Tokens/sec11.456.6

TTFT30.0 s110 ms

Load time1.0 s0.4 s

Memory usage16.5 GB2.2 GB

Memory %52%7%

HW Fit Score Breakdown

deepseek-r1:14b

Speed

20/40

TTFT

0/30

Memory

30/30

phi:2.7b

Speed

40/40

TTFT

30/30

Memory

30/30

Quality

deepseek-r1:14b

Reasoning

1/20

Coding

3/20

Instruction

1/20

Structured

0/15

Math

0/15

Multilingual

5/10

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

phi:2.7b

Reasoning

7/20

Coding

4/20

Instruction

9/20

Structured

11/15

Math

8/15

Multilingual

2/10

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

View deepseek-r1:14b details | View phi:2.7b details

Run yours and compare

npx metrillm@latest bench

Requires Node 20+ and Ollama or LM Studio running