# Liquid AI's LFM2.5 Puts a 24B MoE on Your Device

> The new mixture-of-experts model keeps total capacity high while activating only about 2B parameters per token for edge agents.

Published by The Open Weights on Oct 12, 2026. Canonical: https://theopenweights.com/news/lfm2-5-24b-a2b-15nx

## Key facts

- Company: LiquidAI
- Model: LFM2.5-24B-A2B
- Version: 2.5-24B-A2B
- Category: Embeddings
- Modalities: Text / LLM
- License: Other (Open weights, commercial use allowed)
- Parameters: 24B (mixture of experts)
- Active parameters: 2B
- Context window: 131K tokens
- Architecture: Lfm2MoeForCausalLM
- Precision: BF16
- Weights size: 47.9 GB
- Significance: notable
- Published: 2026-10-12
- Last verified: 2026-10-14
- Hugging Face: https://huggingface.co/LiquidAI/LFM2.5-24B-A2B
- Canonical URL: https://theopenweights.com/news/lfm2-5-24b-a2b-15nx

Liquid AI has released [LFM2.5-24B-A2B](https://huggingface.co/LiquidAI/LFM2.5-24B-A2B), a mixture-of-experts language model designed to run close to the user rather than in a distant data center. The model carries 24 billion total parameters but activates only about 2 billion per token, a design meant to deliver the knowledge capacity of a larger network while keeping the per-inference cost low enough for on-device deployment.

That tradeoff is the whole point. Mixture-of-experts architectures route each token through a small subset of specialized subnetworks, so a model can store far more information than it spends computing on any given step. For edge hardware — laptops, phones, and embedded systems where memory and power are tight — the gap between total and active parameters is what makes a 24B-class model practical at all.

## Why it matters

Most capable open models still assume a GPU server is doing the work. Liquid AI has been building in the opposite direction, aiming its LFM line at local execution and agentic tasks that need to run without a round trip to the cloud.

- **24B total, ~2B active:** broad capability with a small compute footprint per token
- **Multilingual support:** positioned for agents that operate across languages
- **Edge-first:** intended for on-device assistants and automation

As covered in Liquid AI's [launch writeup](https://huggingface.co/blog/LiquidAI/lfm25-24b-a2b), the model targets the growing class of edge agents — software that reasons and acts locally. The release ships under a custom license, so teams planning commercial use should read the terms before building on it.

## Get the model

- [Hugging Face](https://huggingface.co/LiquidAI/LFM2.5-24B-A2B)

## Sources

- [LiquidAI/LFM2.5-24B-A2B](https://huggingface.co/LiquidAI/LFM2.5-24B-A2B) — Hugging Face, Oct 12, 2026

---
Source: The Open Weights (https://theopenweights.com/). Aggregated and written by Claude, curated by humans. Cite as: "Liquid AI's LFM2.5 Puts a 24B MoE on Your Device", The Open Weights, Oct 12, 2026, https://theopenweights.com/news/lfm2-5-24b-a2b-15nx