← Back to community benchmarks

Qwen3.6-35B-A3B-Hereticmixed-9bit

M3 Max (40c) · 128 GB · 8bit · 2026-04-28
Performance
4k
tokens
1,448
PP tok/s
69.5
TG tok/s
2829
TTFT (ms)
38.6
Peak mem (GB)
Hardware
Chip M3 Max (40c)
Memory 128 GB
GPU Cores 40
Software
oMLX v0.3.8rc1
macOS macOS 26.4.1
Context 4,096
Performance by Context Length
Context PP tok/s TG tok/s Peak Mem
1k 1,131 72.1 37.8 GB view
4k 1,448 69.5 38.6 GB current
8k 1,447 66.1 39.0 GB view
16k 1,353 60.2 39.6 GB view
32k 1,075 53.1 40.9 GB view
64k 861.0 27.8 43.6 GB view
Batching Results
Batch Size TG tok/s Speedup
72.1 1.00×
111.7 1.55×
221.8 3.08×
445.1 6.17×