← Back to community benchmarks

Qwen3.5-LLM-122B-A10B-oQ8p

M3 Ultra (80c) · 256 GB · 8bit · 2026-03-31

Performance

16k

tokens

874.7

PP tok/s

36.1

TG tok/s

18732

TTFT (ms)

125.7

Peak mem (GB)

Hardware

Chip M3 Ultra (80c)

Memory 256 GB

GPU Cores 80

Software

oMLX v0.3.0

macOS macOS 26.3.1

Context 16,384

Performance by Context Length

Context	PP tok/s	TG tok/s	Peak Mem
1k	743.6	43.3	122.4 GB	view
4k	912.1	42.0	123.9 GB	view
8k	920.3	39.6	124.5 GB	view
16k	874.7	36.1	125.7 GB	current
32k	766.9	30.2	128.1 GB	view
64k	570.6	23.3	132.9 GB	view
128k	335.1	15.5	142.6 GB	view