← Back to community benchmarks

Gemma-4-26B-A4B-Heretic7bit

M5 Pro (20c) · 48 GB · 7bit · 2026-04-16

Performance

4k

tokens

1,686

PP tok/s

41.5

TG tok/s

2430

TTFT (ms)

22.8

Peak mem (GB)

Hardware

Chip M5 Pro (20c)

Memory 48 GB

GPU Cores 20

Software

oMLX v0.3.5.dev1

macOS macOS 26.3

Context 4,096

Performance by Context Length

Context	PP tok/s	TG tok/s	Peak Mem
1k	1,247	43.4	22.4 GB	view
4k	1,686	41.5	22.8 GB	current

Batching Results

Batch Size	TG tok/s	Speedup
1×	43.4	1.00×
2×	62.2	1.43×
4×	124.5	2.87×