← Back to community benchmarks

SpecPrefill-Qwen3.5-0.8B

M4 Max (40c) · 128 GB · unknown · 2026-04-29

Performance

1k

tokens

3,483

PP tok/s

225.2

TG tok/s

294

TTFT (ms)

2.4

Peak mem (GB)

Hardware

Chip M4 Max (40c)

Memory 128 GB

GPU Cores 40

Software

oMLX v0.3.8rc1

macOS macOS 26.4.1

Context 1,024

Performance by Context Length

Context	PP tok/s	TG tok/s	Peak Mem
1k	3,483	225.2	2.4 GB	current
64k	4,538	104.0	5.4 GB	view
128k	2,729	67.7	8.4 GB	view