whatmodel.sh / LFM2 24B-A2B

LFM2 24B-A2B

+ contribute
Update model informationAdd a place to run itAdd a test resultPropose a benchmarkCorrect a benchmark scoreEdit data on GitHub

LFM2 24B-A2B is an open-weights language model from Liquid AI, released February 2026. Its published size is 23.8B parameters (2.3B active per token) and a 33k-token context window. Licence: lfm1.0.

50.7±1.2#329 of 719 LLMsfrom 11 benchmark results across 1 source

$ how it compares

The LLMs ranked either side of LFM2 24B-A2B, out of 828 with a score, then more from Liquid AI.

#modelscorewhereRAMreleased
396Yi-Large 01.AI51±1comparison only2024-05
397Llama-3-Instruct-8B-SimPO-ExPO Chujie Zheng (ExPO)51±1local~6.6GB2024-05
398LFM2 24B-A2B Liquid AI51±1local~16GB2026-02
399AceReason-Nemotron-1.0-14B NVIDIA51±2local~10GB2025-05
400Llama-3-Instruct-8B-SimPO Princeton NLP51±1local~6.6GB2024-05
more from Liquid AI
370LFM2.5 2.6B Liquid AI53±1local~3.2GB2026-07
375LFM2.5 8B-A1B Liquid AI52±1local~6.5GB2026-05
433LFM2.5 1.2B Thinking Liquid AI49±2local~2.2GB2026-01

see all LLMs · find the ones that fit your machine

$ what it's good at

math▁▂▃▅▆▅▄█▄▁▁▁ #215 of 398 · 31% of its score · 4 tests

Beats DeepSeek V3 (Dec '24), GPT-4o (2024-08-06), Command A (03-2025) here, though they rank higher overall. Trails Olmo 3 7B Instruct, DeepSeek R1 Distill Llama 8B, Phi-4-mini-reasoning 3.8B here, though they rank lower overall.

testresultvs the others in its rundifficultycountswhy
GSM8K87.2% ✎█····█··▄·▄█ beats 4/7very easy · avg 86%▇▇▇▇▇ 13%sharp
MATH (subset unspecified)83.9% ✎█▄··▄█▄····▄ beats 5/7easy · avg 81%▇▇▇ 7.6%typical
PolyMath (Marco evaluation)29.3% ✎▄··▄██··▄··▄ beats 5/7hard · avg 30%▇▇ 5.6%at its level
MGSM (Marco evaluation)67.8% ✎▄·▄·▄▄▄·▄··█ beats 3/7easy · avg 70%▇▇ 5.1%at its level
other tests7 not in a task · 68% of its score
testresultvs the others in its rundifficultycountswhy
GPQA Diamond39.7% ✎▄···█·▄·█·▄▄ beats 5/7hard · avg 38%▇▇▇▇▇▇▇▇ 19%sharp
MMLU74.9% ✎▄···▄▄·▄█··█ beats 3/7easy · avg 75%▇▇▇▇▇▇ 14%at its level
MMLU-Pro57.6% ✎▄····█·▄█·▄▄ beats 5/7medium · avg 56%▇▇▇▇ 11%at its level
MMLU-ProX-Lite (Marco)43.3% ✎█··▄··█··█·▄ beats 4/7medium · avg 43%▇▇▇ 8.6%at its level
GlobalMMLU (Marco evaluation)57.0% ✎▃··▃▃··▅···█ beats 4/7medium · avg 57%▇▇ 6.2%at its level
MMMLU (Marco evaluation)62.3% ✎▄···▄·█·▄·▄█ beats 4/7medium · avg 60%▇▇ 5.4%at its level
AGIEval (Marco suite unspecified)49.0% ✎▄··▄█▄··▄··█ beats 5/7medium · avg 47%▇▇ 4.3%blunt · at its level

$ where to run it

llama.cpp · recommended

version

  1. Install llama.cpp · Homebrew and winget installs name the commands llama-cli and llama-server instead.
    macOS / Linux
    curl -LsSf https://llama.app/install.sh | sh
    Windows
    irm https://llama.app/install.ps1 | iex

    other ways to install ↗

  2. Run LFM2 24B-A2B · needs ~14.4 GB · from huggingface.co/LiquidAI (official)
    chat in the terminal
    llama cli -hf LiquidAI/LFM2-24B-A2B-GGUF:Q4_K_M
  3. Or start a local server and open the address it prints (http://localhost:9931)
    llama serve -hf LiquidAI/LFM2-24B-A2B-GGUF:Q4_K_M

LM Studio

version

  1. Install LM Studio · Open the app once before using the lms command.

    download it (macOS / Windows / Linux) ↗

  2. Run LFM2 24B-A2B · needs ~14.4 GB · from huggingface.co/LiquidAI (official)

    open in LM Studio ↗

  3. Or download it in the terminal
    lms get https://huggingface.co/LiquidAI/LFM2-24B-A2B-GGUF@Q4_K_M -y
  4. Then run lms ls to see its name, and lms chat followed by that name

Ollama

version

  1. Install Ollama

    download it (macOS / Windows) ↗

    Linux
    curl -fsSL https://ollama.com/install.sh | sh
  2. Run LFM2 24B-A2B · needs ~14.78 GB · from ollama.com
    chat in the terminal
    ollama run lfm2:24b

MLX (Mac)

version

  1. Install MLX (Mac) · Apple silicon Macs only, text only.
    with uv
    uv tool install mlx-lm
    or
    pip install mlx-lm
  2. Run LFM2 24B-A2B · needs ~13.4 GB, Mac only · from huggingface.co/LiquidAI (official)
    chat in the terminal
    mlx_lm.chat --model LiquidAI/LFM2-24B-A2B-MLX-4bit
  3. Or start an OpenAI-compatible server at http://localhost:8080/v1
    mlx_lm.server --model LiquidAI/LFM2-24B-A2B-MLX-4bit

Official weights and documentation: huggingface.co/LiquidAI ↗

$ how the score was worked out · Marco-MoE: Open Multilingual… 99%

Each result implies a score on its own, after adjusting for how hard its test is and who ran it. The score is their weighted average, with a small pull toward the middle for thinly tested models. Each test's weight is its counts column above; here it's added up by who ran the tests. How it works

sourceresultsshare of the scorelicence
Marco-MoE: Open Multilingual Mixture-of-Expert Language Models with Efficient Upcycling11▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇ 99%CC BY 4.0
prior · pull toward the middle▇ 1%
weighted average= 50.7

$ improve this page

Update model informationAdd a place to run itAdd a test resultPropose a benchmarkCorrect a benchmark scoreEdit data on GitHub

Use a short GitHub form or edit the data through a pull request. Every number needs a source; here's which ones we can use.

data updated · sources and licences · json