← Back to community benchmarks

Qwen3.5-27B-Hereticmxfp8

M1 Ultra (64c) · 128 GB · 8bit · 2026-03-10
Performance
1k
tokens
163.9
PP tok/s
17.1
TG tok/s
6247
TTFT (ms)
28.4
Peak mem (GB)
Hardware
Chip M1 Ultra (64c)
Memory 128 GB
GPU Cores 64
Software
oMLX v0.2.6
macOS macOS 26.4
Context 1,024
Performance by Context Length
Context PP tok/s TG tok/s Peak Mem
1k 163.9 17.1 28.4 GB current
4k 165.9 16.6 28.7 GB view
8k 164.4 15.8 29.2 GB view
Batching Results
Batch Size TG tok/s Speedup
17.1 1.00×
24.6 1.44×
28.3 1.65×