← Back to community benchmarks

gemma-4-26b-a4b-it

M5 Pro (16c) · 48 GB · 4bit · 2026-04-03
Performance
8k
tokens
1,820
PP tok/s
54.1
TG tok/s
4502
TTFT (ms)
14.9
Peak mem (GB)
Hardware
Chip M5 Pro (16c)
Memory 48 GB
GPU Cores 16
Software
oMLX v0.3.2
macOS macOS 26.3.2
Context 8,192
Performance by Context Length
Context PP tok/s TG tok/s Peak Mem
1k 1,399 61.2 14.3 GB view
4k 1,642 58.6 14.8 GB view
8k 1,820 54.1 14.9 GB current
16k 1,730 42.6 15.3 GB view
Batching Results
Batch Size TG tok/s Speedup
61.2 1.00×
78.2 1.28×
99.1 1.62×