← Back to community benchmarks

gemma4-21b-a4b-REAP-itQ4

M4 (10c) · 32 GB · 4bit · 2026-05-04

Performance

4k

tokens

249.9

PP tok/s

25.2

TG tok/s

16393

TTFT (ms)

12.5

Peak mem (GB)

Hardware

Chip M4 (10c)

Memory 32 GB

GPU Cores 10

Software

oMLX v0.3.8

macOS macOS 15.7.4

Context 4,096

Performance by Context Length

Context	PP tok/s	TG tok/s	Peak Mem
1k	243.8	26.6	11.9 GB	view
4k	249.9	25.2	12.5 GB	current

Batching Results

Batch Size	TG tok/s	Speedup
1×	26.6	1.00×
2×	53.0	1.99×
4×	105.9	3.98×