← Back to community benchmarks

Qwen3.6-35B-A3B

M5 Max (32c) · 36 GB · 4bit · 2026-04-24
Performance
1k
tokens
1,967
PP tok/s
112.8
TG tok/s
521
TTFT (ms)
19.3
Peak mem (GB)
Hardware
Chip M5 Max (32c)
Memory 36 GB
GPU Cores 32
Software
oMLX v0.3.6
macOS macOS 26.4.1
Context 1,024
Performance by Context Length
Context PP tok/s TG tok/s Peak Mem
1k 1,967 112.8 19.3 GB current
4k 3,106 109.0 20.0 GB view
8k 3,281 101.8 20.4 GB view
16k 2,993 92.7 21.0 GB view
Batching Results
Batch Size TG tok/s Speedup
112.8 1.00×
180.5 1.60×
263.0 2.33×
329.3 2.92×