← Back to community benchmarks

Qwen3.6-35B-A3B

M4 Max (40c) · 128 GB · 4bit · 2026-07-23
Performance
16k
tokens
1,421
PP tok/s
94.7
TG tok/s
11534
TTFT (ms)
20.4
Peak mem (GB)
Hardware
Chip M4 Max (40c)
Memory 128 GB
GPU Cores 40
Software
oMLX v0.5.1
macOS macOS 26.5.2
Context 16,384
Performance by Context Length
Context PP tok/s TG tok/s Peak Mem
1k 1,702 108.9 19.3 GB view
4k 1,864 106.4 20.0 GB view
8k 1,714 101.2 20.2 GB view
16k 1,421 94.7 20.4 GB current
32k 1,149 86.5 20.7 GB view
64k 803.5 67.4 21.9 GB view
128k 433.9 35.1 22.9 GB view
195k 320.0 37.9 24.3 GB view
Batching Results
Batch Size TG tok/s Speedup
108.9 1.00×
181.1 1.66×
256.1 2.35×