← Back to community benchmarks

Qwen3.5-35B-A3B-UD-Q4_K_XL

M3 Max (30c) · 36 GB · 3bit · 2026-04-16
Performance
4k
tokens
1,107
PP tok/s
61.3
TG tok/s
3700
TTFT (ms)
22.3
Peak mem (GB)
Hardware
Chip M3 Max (30c)
Memory 36 GB
GPU Cores 30
Software
oMLX v0.3.5
macOS macOS 26.4.1
Context 4,096
Performance by Context Length
Context PP tok/s TG tok/s Peak Mem
1k 883.2 59.2 21.5 GB view
4k 1,107 61.3 22.3 GB current
Batching Results
Batch Size TG tok/s Speedup
59.2 1.00×
101.3 1.71×
141.6 2.39×