LFM3-1.7B · tri-hybrid architecture · 256K context
Device-native intelligence,
streaming live.
Chat with LFM3-1.7B — Liquid AI's 1.7B-parameter model built for local agents, tool calling and structured output. Reasoning traces included, tokens streamed as they generate.
Chat
Engine cold
Parameters
1.73B
Context
256K
Architecture
tri-hybrid
Languages
16
Last run
—