← Back to community benchmarks

gemma-4-31b

M3 Ultra (60c) · 96 GB · 8bit · 2026-04-07

Performance

1k

tokens

239.2

PP tok/s

18.4

TG tok/s

4286

TTFT (ms)

32.1

Peak mem (GB)

Hardware

Chip M3 Ultra (60c)

Memory 96 GB

GPU Cores 60

Software

oMLX v0.3.4

macOS macOS 26.3.1

Context 1,024

Performance by Context Length

Context	PP tok/s	TG tok/s	Peak Mem
1k	239.2	18.4	32.1 GB	current
4k	251.0	17.2	33.6 GB	view

Batching Results

Batch Size	TG tok/s	Speedup
1×	18.4	1.00×
2×	31.1	1.69×
4×	38.5	2.09×