← Back to community benchmarks

LFM2.5-8B-A1B5bit

M4 Pro (16c) · 24 GB · 5bit · 2026-07-17
Performance
16k
tokens
1,385
PP tok/s
114.4
TG tok/s
11827
TTFT (ms)
6.6
Peak mem (GB)
Hardware
Chip M4 Pro (16c)
Memory 24 GB
GPU Cores 16
Software
oMLX v0.5.1
macOS macOS 26.5.2
Context 16,384
Performance by Context Length
Context PP tok/s TG tok/s Peak Mem
1k 1,383 136.8 6.2 GB view
4k 1,506 133.0 6.4 GB view
8k 1,472 126.1 6.5 GB view
16k 1,385 114.4 6.6 GB current
32k 1,248 96.1 6.8 GB view
64k 985.9 72.4 7.3 GB view
Batching Results
Batch Size TG tok/s Speedup
136.8 1.00×
320.8 2.35×