← Back to community benchmarks

Qwen3-Coder-Next

M3 Max (40c) · 128 GB · 6bit · 2026-03-09

Performance

32k

tokens

712.5

PP tok/s

37.8

TG tok/s

45989

TTFT (ms)

63.3

Peak mem (GB)

Hardware

Chip M3 Max (40c)

Memory 128 GB

GPU Cores 40

Software

oMLX v0.2.6

macOS macOS 26.3

Context 32,768

Performance by Context Length

Context	PP tok/s	TG tok/s	Peak Mem
4k	1,033	58.0	61.7 GB	view
8k	1,002	55.5	62.0 GB	view
16k	868.1	50.7	62.4 GB	view
32k	712.5	37.8	63.3 GB	current
64k	578.0	24.1	65.1 GB	view