deepseek-r1:14b

qwen2 · 14.8B · Q4_K_M

Not Rec.

Mar 1, 2026 · Apple M4

deepseek-v2:16b

deepseek2 · 15.7B · Q4_0

Good

Mar 2, 2026 · Apple M4

Global Score

25 vs 73

Hardware Fit

46 vs 100

Quality Score

11 vs 55

Hardware

deepseek-r1:14b deepseek-v2:16b

MachineMacBook AirMacBook Air

CPUApple M4Apple M4

Cores1010

RAM32 GB32 GB

GPUApple M4Apple M4

OSmacOS 26.3macOS 26.3

Archarm64arm64

Power Modeunknownbalanced

Performance

deepseek-r1:14b deepseek-v2:16b

Tokens/sec10.856.6

TTFT30.0 s239 ms

Load time5.3 s1.2 s

Memory usage16.5 GB17.7 GB

Memory %52%55%

HW Fit Score Breakdown

deepseek-r1:14b

Speed

16/40

TTFT

0/30

Memory

30/30

deepseek-v2:16b

Speed

40/40

TTFT

30/30

Memory

30/30

Quality

deepseek-r1:14b

Reasoning

1/20

Coding

0/20

Instruction

2/20

Structured

1/15

Math

1/15

Multilingual

6/10

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

deepseek-v2:16b

Reasoning

7/20

Coding

9/20

Instruction

13/20

Structured

10/15

Math

8/15

Multilingual

8/10

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

View deepseek-r1:14b details | View deepseek-v2:16b details

Run yours and compare

npx metrillm@latest bench

Requires Node 20+ and Ollama or LM Studio running