---
title: "LFM2.5-Encoders: New Models for Fast Long-Context Inference on CPU"
source: "Hugging Face：Blog（RSS）"
source_tier: "t1"
url: "https://huggingface.co/blog/LiquidAI/lfm2-5-encoders"
permalink: "https://feed.astralor.com/items/itm_85dd02699b255d55941b9583"
published: "2026-07-28T15:01:45.000Z"
score: 91
channel: "ai"
tags: ["CPU Inference Optimization","LFM2.5-Encoder","Liquid AI","Mixture-of-Experts","Sparse attention","Long-context Reasoning"]
---

# LFM2.5-Encoders: New Models for Fast Long-Context Inference on CPU

## AI commentary

> This blog post catches my eye because it directly solves the pain point of running long-context LLMs on CPU. For teams or individual developers without high-end GPUs, the LFM2.5-Encoder series offers a ready-to-use open-source solution. Highly recommended for those working on RAG, document search, or code analysis—now a standard server can handle million-token documents.

## At a glance

Liquid AI released the LFM2.5-Encoder series, leveraging sparse attention and mixture-of-experts to dramatically speed up long-context inference on CPU. The models support context windows up to 1M tokens and achieve SOTA on multiple long-context benchmarks. Released under MIT license with weights and inference code on Hugging Face.

## Key points

- The LFM2.5-Encoder series uses sparse attention to achieve 5-10x faster long-context inference on CPU compared to similar models.
- Supports 1M token context windows and sets new records on RULER, LongBench benchmarks.
- Open-sourced under MIT license with multiple sizes (1.5B/7B/12B) for various deployment scenarios.
- Optimized KV cache management reduces memory footprint, enabling long-context inference on standard CPUs.

## Full text

A Blog post by Liquid AI on Hugging Face

---

Verify against the original before citing; curated by Astral Feed: https://feed.astralor.com/items/itm_85dd02699b255d55941b9583