← Back to community benchmarks

Huihui-Qwen3.5-35B-A3B-abliterated

M4 Max (40c) · 48 GB · 6bit · 2026-03-09
Performance
1k
tokens
1,035
PP tok/s
41.6
TG tok/s
989
TTFT (ms)
28.4
Peak mem (GB)
Hardware
Chip M4 Max (40c)
Memory 48 GB
GPU Cores 40
Software
oMLX v0.2.6
macOS macOS 15.4.1
Context 1,024
Performance by Context Length
Context PP tok/s TG tok/s Peak Mem
1k 1,035 41.6 28.4 GB current
4k 1,326 40.4 28.8 GB view
Batching Results
Batch Size TG tok/s Speedup
41.6 1.00×
75.4 1.81×
119.3 2.87×
170.1 4.09×