← Back to community benchmarks

Qwen3.6-35B-A3B

M5 (10c) · 32 GB · 4bit · 2026-04-24
Performance
4k
tokens
1,078
PP tok/s
51.0
TG tok/s
3801
TTFT (ms)
19.9
Peak mem (GB)
Hardware
Chip M5 (10c)
Memory 32 GB
GPU Cores 10
Software
oMLX v0.3.6
macOS macOS 26.3.1
Context 4,096
Performance by Context Length
Context PP tok/s TG tok/s Peak Mem
1k 898.0 52.9 19.2 GB view
4k 1,078 51.0 19.9 GB current
Batching Results
Batch Size TG tok/s Speedup
52.9 1.00×
86.4 1.63×