← Back to community benchmarks

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

M2 Ultra (60c) · 64 GB · 4bit · 2026-07-24
Performance
195k
tokens
100.3
PP tok/s
4.5
TG tok/s
1993510
TTFT (ms)
47.8
Peak mem (GB)
Hardware
Chip M2 Ultra (60c)
Memory 64 GB
GPU Cores 60
Software
oMLX v0.3.5.dev1
macOS macOS 26.3
Context 200,000
Performance by Context Length
Context PP tok/s TG tok/s Peak Mem
1k 175.3 33.1 15.8 GB view
8k 181.6 29.2 17.9 GB view
16k 178.5 26.4 19.1 GB view
32k 171.2 21.3 21.6 GB view
64k 152.8 12.6 26.7 GB view
128k 121.9 6.4 37.0 GB view
195k 100.3 4.5 47.8 GB current
Batching Results
Batch Size TG tok/s Speedup
33.1 1.00×
40.9 1.24×
45.6 1.38×
49.0 1.48×