Back to leaderboard

granite3.1-dense:2b

OLLAMA GGUF

granite · 2.5B · Q4_K_M

Good

Mar 2, 2026 · Apple M4

gemma3:4b

OLLAMA GGUF

gemma3 · 4.3B · Q4_K_M

Good

Jun 26, 2026 · Cortex-A78AE

Global Score
68 vs 76
Hardware Fit
100 vs 92
Quality Score
54 vs 69

Hardware

granite3.1-dense:2b gemma3:4b
MachineMacBook AirNVIDIA NVIDIA Jetson Orin Nano Engineering Reference Developer Kit Super
CPUApple M4Cortex-A78AE
Cores106
RAM32 GB7 GB
GPUApple M4Integrated / Unknown
OSmacOS 26.3Ubuntu 22.04.5 LTS
Archarm64arm64
Power Modebalancedbalanced

Performance

granite3.1-dense:2b gemma3:4b
Tokens/sec54.615.0
First chunkN/A2380 ms
TTFT121 ms2.4 s
Load time0.4 s13.6 s
Memory usage5.6 GB2.6 GB
Memory %18%35%

HW Fit Score Breakdown

granite3.1-dense:2b

Speed
40/50
TTFT
30/20
Memory
30/30

gemma3:4b

Speed
45/50
TTFT
18/20
Memory
29/30

Quality

granite3.1-dense:2b

Reasoning
8/20
Coding
9/20
Instruction
13/20
Structured
14/15
Math
3/15
Multilingual
7/10
Reasoning: Weak Coding: Weak Instruction Following: Adequate Structured Output: Strong Math: Poor Multilingual: Adequate

gemma3:4b

Reasoning
11/20
Coding
17/20
Instruction
13/20
Structured
14/15
Math
4/15
Multilingual
10/10
Reasoning: Adequate Coding: Strong 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