← Back to community benchmarks

Qwen3.5-122B-A10B

M3 Ultra (60c) · 96 GB · 4bit · 2026-03-20
Performance
8k
tokens
748.5
PP tok/s
49.2
TG tok/s
10945
TTFT (ms)
68.5
Peak mem (GB)
Hardware
Chip M3 Ultra (60c)
Memory 96 GB
GPU Cores 60
Software
oMLX v0.2.17
macOS macOS 26.4
Context 8,192
Performance by Context Length
Context PP tok/s TG tok/s Peak Mem
1k 716.3 55.7 66.4 GB view
4k 765.7 52.8 67.9 GB view
8k 748.5 49.2 68.5 GB current
Batching Results
Batch Size TG tok/s Speedup
55.7 1.00×
89.0 1.60×
126.9 2.28×