← Back to community benchmarks

Qwen3.6-35B-A3B-OptiQ

M5 Max (40c) · 128 GB · 4bit · 2026-05-03
Performance
8k
tokens
4,021
PP tok/s
101.9
TG tok/s
2038
TTFT (ms)
21.9
Peak mem (GB)
Hardware
Chip M5 Max (40c)
Memory 128 GB
GPU Cores 40
Software
oMLX v0.3.8
macOS macOS 26.4
Context 8,192
Performance by Context Length
Context PP tok/s TG tok/s Peak Mem
1k 2,280 107.9 20.8 GB view
4k 3,781 104.1 21.6 GB view
8k 4,021 101.9 21.9 GB current
16k 3,621 96.6 22.7 GB view
32k 2,926 87.2 24.2 GB view
64k 2,093 72.4 26.5 GB view
128k 1,269 54.4 31.9 GB view
195k 841.4 41.4 40.1 GB view
Batching Results
Batch Size TG tok/s Speedup
107.9 1.00×
166.9 1.55×
262.6 2.43×