← Back to community benchmarks

LFM2.5-8B-A1B-oQ4

M3 Ultra (60c) · 256 GB · 4bit · 2026-07-23
Performance
64k
tokens
3,472
PP tok/s
157.9
TG tok/s
18875
TTFT (ms)
6.6
Peak mem (GB)
Hardware
Chip M3 Ultra (60c)
Memory 256 GB
GPU Cores 60
Software
oMLX v0.5.3
macOS macOS 26.5.2
Context 65,536
Performance by Context Length
Context PP tok/s TG tok/s Peak Mem
1k 4,686 214.5 5.4 GB view
8k 4,957 206.0 5.6 GB view
16k 4,664 195.5 5.8 GB view
32k 4,193 179.8 6.0 GB view
64k 3,472 157.9 6.6 GB current
128k 2,582 124.2 7.9 GB view
Batching Results
Batch Size TG tok/s Speedup
214.5 1.00×
342.8 1.60×
539.3 2.51×