Liquid AI's post-training team on what makes on-device agentic models actually useful

alexcovo_eth · x · 2026-09-26

Liquid AI's post-training team — Maxime Labonne, Edoard Mosca, and Jiahui Wang — discuss what makes an on-device agentic model actually useful: post-training. They cover how models learn to use tools, follow instructions, handle longer contexts, and recover when tasks get complex.

Related event: Liquid AI releases LFM2.5-2.6B and details its on-device agent post-training recipe(11 posts)→

Original post →

More from coding & agent

coding & agent channel →