← Back to community benchmarks

Josiefied-Qwen3-4B-abliterated-v1

M1 Ultra (48c) · 64 GB · 4bit · 2026-04-28

Performance

4k

tokens

796.6

PP tok/s

75.8

TG tok/s

5142

TTFT (ms)

3.4

Peak mem (GB)

Hardware

Chip M1 Ultra (48c)

Memory 64 GB

GPU Cores 48

Software

oMLX v0.3.6

macOS macOS 26.4.1

Context 4,096

Performance by Context Length

Context	PP tok/s	TG tok/s	Peak Mem
1k	766.9	99.1	2.9 GB	view
4k	796.6	75.8	3.4 GB	current

Batching Results

Batch Size	TG tok/s	Speedup
1×	99.1	1.00×
2×	127.1	1.28×
4×	154.5	1.56×