AI Briefing
KOSign in

LiquidAI releases LFM3-1.7B, a 1GB edge model with 256K context for local agents

·2026.10.28 02:04

Key point

The 1.73B parameter model uses a tri-hybrid architecture to fit on phones and laptops while supporting 256K-token context windows.

Details

  • LiquidAI released LFM3-1.7B, a text model optimized for local agents and specialized applications using a new tri-hybrid architecture.
  • At 4-bit precision, the model requires approximately 1 GB of memory, enabling deployment on smartphones and laptops.
  • Post-trained for agentic harnesses like Hermes Agent, Pi, and OpenClaw, it supports a 256K-token context window and 128,000-token vocabulary.
  • The architecture comprises 24 layers: 13 gated short convolutions, 5 sliding-window attention, and 6 global attention layers.
  • Supports 16 languages including English, Chinese, Japanese, and Korean under the LFM Open License v1.0.
  • Not recommended for agentic coding or complex multi-step reasoning due to its small size; retrieval or search is advised for knowledge-heavy tasks.
open-sourcelocal-llmagentic-aiedge-ailiquidailfm3
Read originalShare on Threads

This summary was generated automatically by AI. Check the original for the author's claims and context. Copyright belongs to the original author.

Our guide explains how the AI works. Report summary errors, attribution issues, or removal requests via Contact.

Related

RedditHarborRL: An Open-Source Framework for Agentic Reinforcement LearningRedditinclusionAI releases Ling-3.1-flash-1019 with 560B parameters and 1M context windowHacker NewsJetBrains releases ThinkRail, an open-source agentic IDE built on git worktrees and spec graphsHacker NewsNvidia OpenShell: Secure Runtime for Autonomous AI Agent FleetsHuggingFace BlogAi2 Releases OlmoEarth v1.3 with State-of-the-Art Global Satellite EmbeddingsHuggingFace BlogIBM releases Granite Speech 5.0 with improved ASR accuracy and 1.6x faster Turbo CTC