LFM2.5-1.2B-Thinking

LFM2.5-1.2B-Thinking is Liquid AI’s compact 1.2B-parameter reasoning model designed for multi-step problem solving, planning, data extraction, and agentic workflows. Its small footprint and 32K context window make it well suited for low-latency inference at the edge, especially for high-volume tasks that need more reasoning than a simple classifier. It is a strong fit for agent planning, routing, tool orchestration, structured extraction, and lightweight reasoning workloads where speed and cost matter. For knowledge-heavy or complex coding tasks, a larger general-purpose model may be a better fit.