← Back to community benchmarks

Qwen3.5-35B-A3B4.8-bit

M1 Max (24c) · 32 GB · 4bit · 2026-03-11

Performance

1k

tokens

394.2

PP tok/s

56.3

TG tok/s

2598

TTFT (ms)

20.4

Peak mem (GB)

Hardware

Chip M1 Max (24c)

Memory 32 GB

GPU Cores 24

Software

oMLX v0.2.7

macOS macOS 15.6

Context 1,024

Performance by Context Length

Context	PP tok/s	TG tok/s	Peak Mem
1k	394.2	56.3	20.4 GB	current
4k	436.8	53.4	20.5 GB	view
8k	435.2	46.3	20.7 GB	view

Batching Results

Batch Size	TG tok/s	Speedup
1×	56.3	1.00×
2×	79.4	1.41×
4×	159.2	2.83×
8×	314.4	5.58×