← Back to community benchmarks

Qwen3.5-122B-A10B

M5 Max (40c) · 128 GB · 4bit · 2026-04-03
Performance
64k
tokens
819.9
PP tok/s
25.4
TG tok/s
79932
TTFT (ms)
76.1
Peak mem (GB)
Hardware
Chip M5 Max (40c)
Memory 128 GB
GPU Cores 40
Software
oMLX v0.3.0
macOS macOS 26.4
Context 65,536
Performance by Context Length
Context PP tok/s TG tok/s Peak Mem
1k 914.2 59.0 65.6 GB view
4k 1,450 59.0 67.0 GB view
8k 1,503 54.3 67.7 GB view
16k 1,357 50.7 68.9 GB view
32k 1,121 39.1 71.2 GB view
64k 819.9 25.4 76.1 GB current
128k 447.9 14.7 85.7 GB view
195k 282.5 10.2 95.9 GB view
Batching Results
Batch Size TG tok/s Speedup
59.0 1.00×
77.3 1.31×
87.5 1.48×
110.3 1.87×