← Back to community benchmarks

gemma-4-e2b-it

M2 Pro (16c) · 32 GB · 4bit · 2026-05-28
Performance
1k
tokens
1,054
PP tok/s
65.8
TG tok/s
971
TTFT (ms)
4.0
Peak mem (GB)
Hardware
Chip M2 Pro (16c)
Memory 32 GB
GPU Cores 16
Software
oMLX v0.3.12
macOS macOS 26.5
Context 1,024
Performance by Context Length
Context PP tok/s TG tok/s Peak Mem
1k 1,054 65.8 4.0 GB current
4k 1,721 68.6 4.0 GB view
8k 1,752 59.6 4.1 GB view
16k 1,637 55.7 4.2 GB view
32k 1,459 46.2 4.5 GB view
64k 1,171 34.5 5.2 GB view
Batching Results
Batch Size TG tok/s Speedup
65.8 1.00×
77.5 1.18×
86.2 1.31×
91.8 1.40×