← Back to community benchmarks

gemma-4-E4B-it

M5 Pro (20c) · 24 GB · 8bit · 2026-05-16
Performance
4k
tokens
3,918
PP tok/s
48.8
TG tok/s
1046
TTFT (ms)
9.1
Peak mem (GB)
Hardware
Chip M5 Pro (20c)
Memory 24 GB
GPU Cores 20
Software
oMLX v0.3.9.dev1
macOS macOS 26.5
Context 4,096
Performance by Context Length
Context PP tok/s TG tok/s Peak Mem
1k 3,588 49.9 8.9 GB view
4k 3,918 48.8 9.1 GB current
8k 3,925 47.3 9.2 GB view
16k 3,774 44.3 9.5 GB view
32k 3,437 39.9 10.2 GB view
64k 2,866 33.1 11.0 GB view
128k 2,041 24.7 13.3 GB view
Batching Results
Batch Size TG tok/s Speedup
49.9 1.00×
86.8 1.74×
124.5 2.49×
256.4 5.14×