Liquid AI's New LFM 2.5 Is INSANE — 184 TOK/S With Only 2B Active! (Live Tests) - High-performance local edge reasoning guide, prompts, and deployment for LFM2.5-24B-A2B hybrid architecture