← Back to community benchmarks

Qwen3.6-35B-A3B

M3 Max (40c) · 128 GB · 6bit · 2026-04-26
Performance
16k
tokens
1,024
PP tok/s
57.4
TG tok/s
16000
TTFT (ms)
29.1
Peak mem (GB)
Hardware
Chip M3 Max (40c)
Memory 128 GB
GPU Cores 40
Software
oMLX v0.3.6
macOS macOS 26.4.1
Context 16,384
Performance by Context Length
Context PP tok/s TG tok/s Peak Mem
1k 1,006 74.3 27.3 GB view
4k 1,393 71.1 28.1 GB view
16k 1,024 57.4 29.1 GB current
32k 852.7 50.8 30.4 GB view
64k 622.6 28.2 33.1 GB view
128k 230.2 15.7 38.5 GB view
195k 186.6 9.5 44.1 GB view
Batching Results
Batch Size TG tok/s Speedup
74.3 1.00×
57.7 0.78×
63.1 0.85×
70.8 0.95×