← Back to community benchmarks

Qwen3.5-0.8B

M5 (10c) · 24 GB · 8bit · 2026-03-10
Performance
64k
tokens
1,217
PP tok/s
61.2
TG tok/s
53850
TTFT (ms)
4.4
Peak mem (GB)
Hardware
Chip M5 (10c)
Memory 24 GB
GPU Cores 10
Software
oMLX v0.2.7
macOS macOS 26.3
Context 65,536
Performance by Context Length
Context PP tok/s TG tok/s Peak Mem
1k 2,530 134.0 1.8 GB view
4k 2,505 125.4 1.9 GB view
8k 2,333 104.7 2.0 GB view
16k 2,098 71.2 2.4 GB view
32k 1,704 54.4 3.0 GB view
64k 1,217 61.2 4.4 GB current
Batching Results
Batch Size TG tok/s Speedup
134.0 1.00×
515.6 3.85×