← Back to community benchmarks

Qwen3.6-35B-A3B-uncensored-heretic-Native-MTP-Preserved

M5 Max (40c) · 128 GB · fp16 · 2026-07-25
Performance
195k
tokens
810.8
PP tok/s
38.5
TG tok/s
246663
TTFT (ms)
82.4
Peak mem (GB)
Hardware
Chip M5 Max (40c)
Memory 128 GB
GPU Cores 40
Software
oMLX v0.5.3
macOS macOS 27.0
Context 200,000
Performance by Context Length
Context PP tok/s TG tok/s Peak Mem
1k 1,950 69.4 65.8 GB view
4k 2,621 69.2 66.6 GB view
8k 3,780 68.2 66.9 GB view
16k 3,167 66.0 67.6 GB view
32k 2,794 62.2 68.9 GB view
64k 2,068 55.7 71.5 GB view
128k 1,275 47.4 76.9 GB view
195k 810.8 38.5 82.4 GB current
Batching Results
Batch Size TG tok/s Speedup
69.4 1.00×
65.9 0.95×
116.2 1.67×