← Back to community benchmarks

Qwen3.6-35B-A3B-UD

M4 Max (40c) · 128 GB · 4bit · 2026-04-29
Performance
16k
tokens
1,586
PP tok/s
76.2
TG tok/s
10328
TTFT (ms)
22.2
Peak mem (GB)
Hardware
Chip M4 Max (40c)
Memory 128 GB
GPU Cores 40
Software
oMLX v0.3.8rc1
macOS macOS 26.3
Context 16,384
Performance by Context Length
Context PP tok/s TG tok/s Peak Mem
1k 1,219 93.2 20.4 GB view
4k 1,642 88.6 21.2 GB view
8k 1,685 82.5 21.5 GB view
16k 1,586 76.2 22.2 GB current
Batching Results
Batch Size TG tok/s Speedup
93.2 1.00×
152.4 1.64×
233.1 2.50×
298.7 3.20×