← Back to community benchmarks

Qwen3.5-122B-A10B

M3 Ultra (80c) · 512 GB · 8bit · 2026-03-22
Performance
32k
tokens
766.6
PP tok/s
28.4
TG tok/s
42745
TTFT (ms)
128.9
Peak mem (GB)
Hardware
Chip M3 Ultra (80c)
Memory 512 GB
GPU Cores 80
Software
oMLX v0.2.18
macOS macOS 26.2
Context 32,768
Performance by Context Length
Context PP tok/s TG tok/s Peak Mem
1k 861.8 38.7 123.3 GB view
4k 956.1 37.3 124.8 GB view
8k 940.6 35.8 125.4 GB view
16k 881.2 33.2 126.5 GB view
32k 766.6 28.4 128.9 GB current
Batching Results
Batch Size TG tok/s Speedup
38.7 1.00×
63.4 1.64×
92.9 2.40×
132.7 3.43×