← Back to community benchmarks

LFM2-24B-A2B

M5 (10c) · 24 GB · 4bit · 2026-03-10
Performance
32k
tokens
523.3
PP tok/s
47.6
TG tok/s
62622
TTFT (ms)
13.6
Peak mem (GB)
Hardware
Chip M5 (10c)
Memory 24 GB
GPU Cores 10
Software
oMLX v0.2.7
macOS macOS 26.3.1
Context 32,768
Performance by Context Length
Context PP tok/s TG tok/s Peak Mem
1k 625.8 71.3 13.0 GB view
4k 603.2 72.2 13.1 GB view
8k 614.2 65.6 13.1 GB view
16k 583.7 59.6 13.3 GB view
32k 523.3 47.6 13.6 GB current
Batching Results
Batch Size TG tok/s Speedup
71.3 1.00×
106.2 1.49×
124.0 1.74×