← Back to community benchmarks

Qwen3.5-0.8B

M2 Ultra (76c) · 192 GB · 4bit · 2026-03-15
Performance
64k
tokens
3,930
PP tok/s
67.3
TG tok/s
16676
TTFT (ms)
4.1
Peak mem (GB)
Hardware
Chip M2 Ultra (76c)
Memory 192 GB
GPU Cores 76
Software
oMLX v0.2.12
macOS macOS 15.0
Context 65,536
Performance by Context Length
Context PP tok/s TG tok/s Peak Mem
64k 3,930 67.3 4.1 GB current
128k 2,470 29.3 7.0 GB view
195k 1,715 32.9 10.1 GB view