← Back to community benchmarks

Qwen3.5-27B-heretic

M1 Ultra (64c) · 128 GB · 8bit · 2026-03-10
Performance
8k
tokens
160.1
PP tok/s
16.6
TG tok/s
51161
TTFT (ms)
29.9
Peak mem (GB)
Hardware
Chip M1 Ultra (64c)
Memory 128 GB
GPU Cores 64
Software
oMLX v0.2.6
macOS macOS 26.4
Context 8,192
Performance by Context Length
Context PP tok/s TG tok/s Peak Mem
1k 161.8 17.6 29.1 GB view
4k 161.1 17.6 29.5 GB view
8k 160.1 16.6 29.9 GB current
16k 156.7 15.1 30.8 GB view
Batching Results
Batch Size TG tok/s Speedup
17.6 1.00×
31.4 1.78×
38.3 2.18×
42.1 2.39×