← Back to community benchmarks

Llama-4-Scout-17B-16E-Instruct

M3 Ultra (60c) · 256 GB · 6bit · 2026-05-20

Performance

1k

tokens

454.5

PP tok/s

32.0

TG tok/s

2255

TTFT (ms)

82.2

Peak mem (GB)

Hardware

Chip M3 Ultra (60c)

Memory 256 GB

GPU Cores 60

Software

oMLX v0.3.9.dev2

macOS macOS 26.5

Context 1,024

Performance by Context Length

Context	PP tok/s	TG tok/s	Peak Mem
1k	454.5	32.0	82.2 GB	current
4k	496.1	30.0	83.2 GB	view

Batching Results

Batch Size	TG tok/s	Speedup
1×	32.0	1.00×
2×	64.2	2.01×
4×	127.6	3.99×