← Back to community benchmarks

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

M3 Ultra (80c) · 512 GB · 4bit · 2026-03-09
Performance
16k
tokens
1,961
PP tok/s
64.3
TG tok/s
8355
TTFT (ms)
20.8
Peak mem (GB)
Hardware
Chip M3 Ultra (80c)
Memory 512 GB
GPU Cores 80
Software
oMLX v0.2.6
macOS macOS 26.3
Context 16,384
Performance by Context Length
Context PP tok/s TG tok/s Peak Mem
1k 1,540 83.5 19.2 GB view
4k 2,016 80.7 19.5 GB view
8k 2,069 73.4 20.0 GB view
16k 1,961 64.3 20.8 GB current
32k 1,650 53.0 22.4 GB view
64k 1,136 38.1 25.7 GB view
Batching Results
Batch Size TG tok/s Speedup
83.5 1.00×
138.3 1.66×
239.0 2.86×
361.7 4.33×