← Back to community benchmarks

supergemma4-26b-uncensored4bit-v2

M5 Pro (20c) · 64 GB · 4bit · 2026-04-15

Performance

4k

tokens

1,590

PP tok/s

72.2

TG tok/s

2577

TTFT (ms)

14.4

Peak mem (GB)

Hardware

Chip M5 Pro (20c)

Memory 64 GB

GPU Cores 20

Software

oMLX v0.3.5

macOS macOS 26.4.1

Context 4,096

Performance by Context Length

Context	PP tok/s	TG tok/s	Peak Mem
1k	1,149	76.4	14.0 GB	view
4k	1,590	72.2	14.4 GB	current

Batching Results

Batch Size	TG tok/s	Speedup
1×	76.4	1.00×
2×	113.1	1.48×
4×	140.7	1.84×