← Back to community benchmarks

Qwen3.5-35B-A3B

M5 Max (32c) · 36 GB · 4bit · 2026-03-15
User commentrun in low power mode
Performance
4k
tokens
1,279
PP tok/s
53.5
TG tok/s
3203
TTFT (ms)
20.2
Peak mem (GB)
Hardware
Chip M5 Max (32c)
Memory 36 GB
GPU Cores 32
Software
oMLX v0.2.12
macOS macOS 26.3.1
Context 4,096
Performance by Context Length
Context PP tok/s TG tok/s Peak Mem
1k 2,352 58.9 20.0 GB view
4k 1,279 53.5 20.2 GB current
Batching Results
Batch Size TG tok/s Speedup
58.9 1.00×
91.5 1.55×
120.6 2.05×