← Back to community benchmarks

DeepSeek-R1-0528-Qwen3-8B-4bit-AWQ

M4 (10c) · 16 GB · 4bit · 2026-03-18

Performance

1k

tokens

221.1

PP tok/s

21.9

TG tok/s

4631

TTFT (ms)

5.0

Peak mem (GB)

Hardware

Chip M4 (10c)

Memory 16 GB

GPU Cores 10

Software

oMLX v0.2.18

macOS macOS 26.3.1

Context 1,024

Performance by Context Length

Context	PP tok/s	TG tok/s	Peak Mem
1k	221.1	21.9	5.0 GB	current
4k	208.5	19.4	5.5 GB	view

Batching Results

Batch Size	TG tok/s	Speedup
1×	21.9	1.00×
2×	42.7	1.95×
4×	53.0	2.42×