← Back to community benchmarks

gemma-3-4b-it-qat

M4 (10c) · 24 GB · 4bit · 2026-05-13
Performance
195k
tokens
171.0
PP tok/s
11.8
TG tok/s
1169342
TTFT (ms)
10.7
Peak mem (GB)
Hardware
Chip M4 (10c)
Memory 24 GB
GPU Cores 10
Software
oMLX v0.3.8
macOS macOS 26.3.1
Context 200,000
Performance by Context Length
Context PP tok/s TG tok/s Peak Mem
1k 160.0 18.4 3.1 GB view
4k 169.7 17.5 3.5 GB view
8k 161.2 16.3 3.6 GB view
16k 142.6 14.9 4.0 GB view
32k 154.3 22.2 4.7 GB view
64k 290.0 20.2 5.4 GB view
128k 213.9 12.2 8.0 GB view
195k 171.0 11.8 10.7 GB current
Batching Results
Batch Size TG tok/s Speedup
18.4 1.00×
60.8 3.30×
67.1 3.65×
70.0 3.80×