← Back to community benchmarks

Qwen3.5-27B-Claude-4.6-Opus-Distilled

M3 Max (40c) · 64 GB · 4bit · 2026-03-20
Performance
195k
tokens
109.7
PP tok/s
4.3
TG tok/s
1822666
TTFT (ms)
47.9
Peak mem (GB)
Hardware
Chip M3 Max (40c)
Memory 64 GB
GPU Cores 40
Software
oMLX v0.2.19
macOS macOS 26.3.1
Context 200,000
Performance by Context Length
Context PP tok/s TG tok/s Peak Mem
1k 209.4 23.1 15.8 GB view
4k 216.4 22.1 17.4 GB view
8k 197.8 21.4 18.0 GB view
16k 202.2 19.5 19.3 GB view
32k 191.1 17.2 21.8 GB view
64k 166.3 12.9 26.9 GB view
128k 133.7 6.0 37.2 GB view
195k 109.7 4.3 47.9 GB current
Batching Results
Batch Size TG tok/s Speedup
23.1 1.00×
38.5 1.67×
46.8 2.03×
51.9 2.25×