← Back to community benchmarks

gemma-4-E2B-it-qat

M5 (10c) · 24 GB · 4bit · 2026-06-05
Performance
16k
tokens
4,770
PP tok/s
47.1
TG tok/s
3435
TTFT (ms)
4.0
Peak mem (GB)
Hardware
Chip M5 (10c)
Memory 24 GB
GPU Cores 10
Software
oMLX v0.4.2.dev3
macOS macOS 26.5.1
Context 16,384
Performance by Context Length
Context PP tok/s TG tok/s Peak Mem
1k 3,546 49.9 3.7 GB view
4k 4,909 50.8 3.7 GB view
16k 4,770 47.1 4.0 GB current
32k 4,213 42.1 4.4 GB view
64k 3,232 35.0 5.3 GB view
128k 2,161 25.1 6.7 GB view
195k 1,578 17.7 9.2 GB view
Batching Results
Batch Size TG tok/s Speedup
49.9 1.00×
88.6 1.78×
117.8 2.36×
141.8 2.84×