← Back to community benchmarks

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

M5 Pro (20c) · 48 GB · 8bit · 2026-04-25

Performance

1k

tokens

1,209

PP tok/s

65.6

TG tok/s

847

TTFT (ms)

35.4

Peak mem (GB)

Hardware

Chip M5 Pro (20c)

Memory 48 GB

GPU Cores 20

Software

oMLX v0.3.6

macOS macOS 26.4.1

Context 1,024

Performance by Context Length

Context	PP tok/s	TG tok/s	Peak Mem
1k	1,209	65.6	35.4 GB	current
4k	1,995	64.6	36.2 GB	view

Batching Results

Batch Size	TG tok/s	Speedup
1×	65.6	1.00×
2×	107.9	1.64×
4×	137.6	2.10×