← Back to community benchmarks

Qwen3.6-35B-A3B-UD-Q2_K_XL

M5 (10c) · 24 GB · 2bit · 2026-05-20
Performance
4k
tokens
1,080
PP tok/s
50.5
TG tok/s
3791
TTFT (ms)
14.1
Peak mem (GB)
Hardware
Chip M5 (10c)
Memory 24 GB
GPU Cores 10
Software
oMLX v0.3.9.dev2
macOS macOS 26.5
Context 4,096
Performance by Context Length
Context PP tok/s TG tok/s Peak Mem
1k 857.9 52.2 13.4 GB view
4k 1,080 50.5 14.1 GB current
8k 1,066 47.9 14.6 GB view
16k 976.1 43.9 15.3 GB view
32k 798.6 36.9 16.8 GB view
Batching Results
Batch Size TG tok/s Speedup
52.2 1.00×
68.1 1.30×
89.3 1.71×
177.3 3.40×