← Back to community benchmarks

Qwen3.5-122B-A10B

M3 Ultra (80c) · 512 GB · 8bit · 2026-04-03
Performance
4k
tokens
915.4
PP tok/s
40.1
TG tok/s
4475
TTFT (ms)
123.9
Peak mem (GB)
Hardware
Chip M3 Ultra (80c)
Memory 512 GB
GPU Cores 80
Software
oMLX v0.3.0
macOS macOS 26.2
Context 4,096
Performance by Context Length
Context PP tok/s TG tok/s Peak Mem
1k 742.0 41.5 122.5 GB view
4k 915.4 40.1 123.9 GB current
8k 919.9 38.4 124.5 GB view
16k 874.8 35.3 125.7 GB view
Batching Results
Batch Size TG tok/s Speedup
41.5 1.00×
66.0 1.59×
90.3 2.18×
120.7 2.91×