← Back to community benchmarks

gemma-4-26B-A4B-it-QAT

M5 Pro (20c) · 64 GB · 4bit · 2026-07-06

Performance

32k

tokens

1,774

PP tok/s

52.8

TG tok/s

18471

TTFT (ms)

16.3

Peak mem (GB)

Hardware

Chip M5 Pro (20c)

Memory 64 GB

GPU Cores 20

Software

oMLX v0.4.4

macOS macOS 26.5.2

Context 32,768

Performance by Context Length

Context	PP tok/s	TG tok/s	Peak Mem
1k	1,506	80.6	14.3 GB	view
8k	2,022	72.1	15.0 GB	view
32k	1,774	52.8	16.3 GB	current