← Back to community benchmarks

Qwen3.5-35B-A3B5.5bit

M4 (10c) · 32 GB · 5bit · 2026-03-10
Performance
1k
tokens
396.1
PP tok/s
40.5
TG tok/s
2586
TTFT (ms)
23.2
Peak mem (GB)
Hardware
Chip M4 (10c)
Memory 32 GB
GPU Cores 10
Software
oMLX v0.2.6
macOS macOS 26.3.1
Context 1,024
Performance by Context Length
Context PP tok/s TG tok/s Peak Mem
1k 396.1 40.5 23.2 GB current
4k 432.5 38.9 23.4 GB view
8k 413.0 36.5 23.6 GB view
16k 369.3 31.6 24.0 GB view