← Back to community benchmarks

gemma-4-E4B-it-Claude-Opus-4.5-HERETIC-UNCENSORED-Thinking-oQ3

M2 Pro (16c) · 16 GB · 3bit · 2026-04-29

Performance

4k

tokens

807.9

PP tok/s

51.0

TG tok/s

5070

TTFT (ms)

4.0

Peak mem (GB)

Hardware

Chip M2 Pro (16c)

Memory 16 GB

GPU Cores 16

Software

oMLX v0.3.6

macOS macOS 26.4.1

Context 4,096

Performance by Context Length

Context	PP tok/s	TG tok/s	Peak Mem
1k	696.2	54.1	3.9 GB	view
4k	807.9	51.0	4.0 GB	current

Batching Results

Batch Size	TG tok/s	Speedup
1×	54.1	1.00×
2×	66.4	1.23×
4×	73.4	1.36×