← Back to community benchmarks

Qwen3.5-35B-A3B-5bit

M3 Ultra (60c) · 96 GB · 5bit · 2026-03-12

Performance

4k

tokens

2,091

PP tok/s

84.2

TG tok/s

1959

TTFT (ms)

24.2

Peak mem (GB)

Hardware

Chip M3 Ultra (60c)

Memory 96 GB

GPU Cores 60

Software

oMLX v0.2.8

macOS macOS 26.2

Context 4,096

Performance by Context Length

Context	PP tok/s	TG tok/s	Peak Mem
1k	2,010	86.9	24.1 GB	view
4k	2,091	84.2	24.2 GB	current

Batching Results

Batch Size	TG tok/s	Speedup
1×	86.9	1.00×
2×	143.3	1.65×
4×	284.8	3.28×