← Back to community benchmarks

Qwen3.5-122B-A10B

M3 Max (30c) · 96 GB · 4bit · 2026-04-04

Performance

16k

tokens

1,081

PP tok/s

9.1

TG tok/s

15150

TTFT (ms)

69.6

Peak mem (GB)

Hardware

Chip M3 Max (30c)

Memory 96 GB

GPU Cores 30

Software

oMLX v0.3.2

macOS macOS 26.4

Context 16,384

Performance by Context Length

Context	PP tok/s	TG tok/s	Peak Mem
1k	315.7	34.5	66.4 GB	view
4k	346.1	31.9	68.2 GB	view
8k	327.5	28.3	68.8 GB	view
16k	1,081	9.1	69.6 GB	current
32k	1,075	9.1	71.0 GB	view
64k	952.5	9.0	72.8 GB	view
128k	711.5	9.1	76.3 GB	view