← Back to community benchmarks

Huihui-Qwen3.6-35B-A3B-Claude-4.7-Opus-abliterated

M5 Max (40c) · 128 GB · 8bit · 2026-05-18
Performance
16k
tokens
3,624
PP tok/s
90.5
TG tok/s
4521
TTFT (ms)
37.4
Peak mem (GB)
Hardware
Chip M5 Max (40c)
Memory 128 GB
GPU Cores 40
Software
oMLX v0.3.8
macOS macOS 26.4.1
Context 16,384
Performance by Context Length
Context PP tok/s TG tok/s Peak Mem
1k 1,476 98.1 35.4 GB view
4k 3,192 91.6 36.2 GB view
8k 3,843 95.1 36.7 GB view
16k 3,624 90.5 37.4 GB current
32k 3,155 84.0 38.9 GB view
64k 2,248 70.5 41.9 GB view
128k 1,268 49.0 47.9 GB view
Batching Results
Batch Size TG tok/s Speedup
98.1 1.00×
145.5 1.48×
208.5 2.13×
249.6 2.54×