Back to leaderboard

deepseek-v2:16b

deepseek2 · 15.7B · Q4_0

Good

Mar 2, 2026 · Apple M4

qwen3:8b

qwen3 · 8.2B · Q4_K_M

Marginal

Mar 1, 2026 · Apple M4

Global Score
73 vs 42
Hardware Fit
100 vs 60
Quality Score
55 vs 30

Hardware

deepseek-v2:16b qwen3:8b
MachineMacBook AirMacBook Air
CPUApple M4Apple M4
Cores1010
RAM32 GB32 GB
GPUApple M4Apple M4
OSmacOS 26.3macOS 26.3
Archarm64arm64
Power Modebalancedunknown

Performance

deepseek-v2:16b qwen3:8b
Tokens/sec56.618.2
TTFT239 ms12.4 s
Load time1.2 s2.7 s
Memory usage17.7 GB9.5 GB
Memory %55%30%

HW Fit Score Breakdown

deepseek-v2:16b

Speed
40/40
TTFT
30/30
Memory
30/30

qwen3:8b

Speed
27/40
TTFT
3/30
Memory
30/30

Quality

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

qwen3:8b

Reasoning
9/20
Coding
0/20
Instruction
5/20
Structured
4/15
Math
3/15
Multilingual
9/10
Reasoning: Weak Coding: Poor Instruction Following: Weak Structured Output: Weak Math: Poor Multilingual: Strong
View deepseek-v2:16b details | View qwen3:8b details

Run yours and compare

npx metrillm@latest bench

Requires Node 20+ and Ollama or LM Studio running