← Back to community benchmarks

Qwen3___5-397B-A17B

M3 Ultra (80c) · 512 GB · 4bit · 2026-03-15

Performance

32k

tokens

446.5

PP tok/s

21.1

TG tok/s

73396

TTFT (ms)

213.2

Peak mem (GB)

Hardware

Chip M3 Ultra (80c)

Memory 512 GB

GPU Cores 80

Software

oMLX v0.2.12

macOS macOS 15.3

Context 32,768