Liquid AI/Mixture of Experts

LFM2 24B A2B

Name: LFM2 24B A2B
Author: Liquid AI

LFM2 is a family of hybrid models designed for on-device deployment. LFM2-24B-A2B is the largest model in the family, scaling the architecture to 24 billion parameters while keeping inference efficient.

chatreasoningtool_usemultilingual

24B

Parameters (2B active)

125K

Context length

Benchmarks

Quantizations

Architecture

MoE

Released

2026-02-12

Layers

KV Heads

Head Dim

Family

lfm

Quantization Options

Quant	Bits	VRAM	Quality
Q3_K_M	4	12.5 GB	low
Q3_K_L	4.3	13.4 GB	moderate
IQ4_XS	4.46	13.9 GB	moderate
Q4_K_S	4.67	14.5 GB	moderate
Q4_K_M	4.89	15.2 GB	good
Q5_K_S	5.57	17.2 GB	good
Q5_K_M	5.7	17.6 GB	good
Q6_K	6.56	20.2 GB	excellent
Q8_0	8.5	26.0 GB	lossless
FP16	16	48.5 GB	lossless

Select your GPU above to see speed estimates and compatibility for each quantization.

▸ READY TO RUN THIS?RENT BY THE HOUR

RENT A GPU AND RUN LFM2 24B A2B NOW

Rent on RunPod →Or Vast.ai →

Spin up an A100 / H100 / 4090 in ~60s. Pay by the second. Cancel anytime.

Community Ratings

Loading ratings...

Benchmarks (7)

GPQA Diamond47.4

IFBench45.9

τ²-Bench11.1

SciCode10.9

AA Intelligence10.5

HLE4.4

AA Coding3.6

Run this model

▸Easiest way to get started·Beginners

DOCS ↗

curl -fsSL https://ollama.com/install.sh | sh

$ollama run lfm:24b-q4_K_M

Tag may need adjustment — check ollama.com/library/lfm for available tags.

▸ SETUP GUIDE

Auto-setup with fitmyllm CLI

Detects your GPU, recommends the best model, downloads it, and starts chatting — zero config. Benchmarks your speed and contributes anonymous data to improve predictions.

pip install fitmyllmthen run fitmyllmLearn more

Auto-detect GPULive tok/s in chatSpeed benchmarks9 inference engines

HuggingFace GGUF Downloads Build Hardware

GPUs that can run this model

At Q4_K_M quantization. Sorted by minimum VRAM.

NVIDIA RTX 5080

16 GB VRAM • 960 GB/s

NVIDIA

$999