← Back to community benchmarks

Qwen3.5-122B-A10B

M5 Max (40c) · 128 GB · 4bit · 2026-04-30
User commentoMLX v0.3.8
Performance
16k
tokens
1,451
PP tok/s
58.1
TG tok/s
11289
TTFT (ms)
69.0
Peak mem (GB)
Hardware
Chip M5 Max (40c)
Memory 128 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 963.6 66.2 65.6 GB view
4k 1,532 64.5 66.9 GB view
8k 1,581 61.9 67.8 GB view
16k 1,451 58.1 69.0 GB current
32k 1,223 52.6 71.6 GB view
64k 923.4 43.6 76.8 GB view
128k 611.2 33.5 87.2 GB view
195k 446.0 25.9 98.8 GB view
Batching Results
Batch Size TG tok/s Speedup
66.2 1.00×
98.3 1.48×
135.0 2.04×