← Back to community benchmarks

Huihui-Qwen3.5-35B-A3B-abliterated-4bit

M1 Max (32c) · 32 GB · 4bit · 2026-03-11
Performance
8k
tokens
583.1
PP tok/s
56.9
TG tok/s
14050
TTFT (ms)
20.4
Peak mem (GB)
Hardware
Chip M1 Max (32c)
Memory 32 GB
GPU Cores 32
Software
oMLX v0.2.8
macOS macOS 26.3.1
Context 8,192
Performance by Context Length
Context PP tok/s TG tok/s Peak Mem
1k 572.7 63.4 20.1 GB view
4k 594.2 60.2 20.2 GB view
8k 583.1 56.9 20.4 GB current
16k 559.9 46.6 20.8 GB view
32k 512.8 34.7 21.6 GB view
64k 416.9 23.2 23.3 GB view
Batching Results
Batch Size TG tok/s Speedup
63.4 1.00×
92.2 1.45×
117.3 1.85×
131.3 2.07×