← Back to community benchmarks

qw3.6-35b-a3b-4bit-dwq

M5 Max (40c) · 64 GB · 8bit · 2026-05-12
Performance
16k
tokens
3,670
PP tok/s
103.2
TG tok/s
4464
TTFT (ms)
22.3
Peak mem (GB)
Hardware
Chip M5 Max (40c)
Memory 64 GB
GPU Cores 40
Software
oMLX v0.3.8
macOS macOS 26.4.1
Context 16,384
Performance by Context Length
Context PP tok/s TG tok/s Peak Mem
1k 2,333 117.5 20.3 GB view
4k 3,826 113.1 21.1 GB view
8k 4,074 107.9 21.5 GB view
16k 3,670 103.2 22.3 GB current
32k 3,012 94.8 23.8 GB view
64k 2,177 77.3 26.8 GB view
128k 1,283 53.1 32.8 GB view
Batching Results
Batch Size TG tok/s Speedup
117.5 1.00×
167.6 1.43×