← Back to community benchmarks

gemma-4-E4B-Opus-Fused-94kf

M5 Pro (20c) · 64 GB · 4bit · 2026-04-17
Performance
4k
tokens
2,948
PP tok/s
72.3
TG tok/s
1389
TTFT (ms)
28.3
Peak mem (GB)
Hardware
Chip M5 Pro (20c)
Memory 64 GB
GPU Cores 20
Software
oMLX v0.3.6
macOS macOS 26.4.1
Context 4,096
Performance by Context Length
Context PP tok/s TG tok/s Peak Mem
1k 1,752 61.7 28.2 GB view
4k 2,948 72.3 28.3 GB current
Batching Results
Batch Size TG tok/s Speedup
61.7 1.00×
147.0 2.38×
180.2 2.92×