← Back to community benchmarks

gemma-4-e4b-it

M4 (10c) · 24 GB · 4bit · 2026-05-23
Performance
195k
tokens
277.3
PP tok/s
10.7
TG tok/s
721121
TTFT (ms)
12.6
Peak mem (GB)
Hardware
Chip M4 (10c)
Memory 24 GB
GPU Cores 10
Software
oMLX v0.3.9
macOS macOS 26.5
Context 200,000
Performance by Context Length
Context PP tok/s TG tok/s Peak Mem
1k 575.2 28.4 5.4 GB view
4k 612.7 28.8 5.5 GB view
8k 578.9 26.4 5.7 GB view
16k 540.1 25.7 5.9 GB view
32k 518.8 24.8 6.5 GB view
64k 462.6 19.8 7.6 GB view
128k 353.1 14.3 9.8 GB view
195k 277.3 10.7 12.6 GB current
Batching Results
Batch Size TG tok/s Speedup
28.4 1.00×
61.3 2.16×
83.0 2.92×
88.4 3.11×