LLM EXPLORER 65,418 MODELS INDEXED

LFM2.5 24B A2B MLX 8bit by LiquidAI

By LiquidAI · 2366 downloads
HF Repo ↗ Copy URL

LFM2.5 24B A2B MLX 8bit is an open-source language model by LiquidAI. Features: 24b LLM, VRAM: 25.6GB, Context: 128K, License: other, Quantized, LLM Explorer Score: 0.35.

  Arxiv:2511.23404   8-bit   8bit   Ar Base model:liquidai/lfm2.5-24b... Base model:quantized:liquidai/...   Conversational   De   Edge   En   Es   Fr   It   Ja   Ko   Lfm2.5   Lfm2 moe   Liquid   Mlx   Moe   Pt   Quantized   Region:us   Safetensors   Sharded   Tensorflow   Zh
Model Card on HF 🤗: https://huggingface.co/LiquidAI/LFM2.5-24B-A2B-MLX-8bit 

LFM2.5 24B A2B MLX 8bit Parameters and Internals

LLM NameLFM2.5 24B A2B MLX 8bit
Repository 🤗https://huggingface.co/LiquidAI/LFM2.5-24B-A2B-MLX-8bit 
Base Model(s)  LiquidAI/LFM2.5-24B-A2B   LiquidAI/LFM2.5-24B-A2B
Model Size24b
Required VRAM25.6 GB
Updated2026-10-22
MaintainerLiquidAI
Model Typelfm2_moe
Model Files  5.3 GB: 1-of-5   5.3 GB: 2-of-5   5.3 GB: 3-of-5   5.3 GB: 4-of-5   4.4 GB: 5-of-5
Supported Languagesen ar zh fr de ja ko es pt it
Quantization Type8bit
Model ArchitectureLfm2MoeForCausalLM
Licenseother
Context Length131072
Model Max Length131072
Transformers Version5.12.1
Tokenizer ClassTokenizersBackend
Padding Token<|pad|>
Vocabulary Size128000

Best Alternatives to LFM2.5 24B A2B MLX 8bit

Best Alternatives
Context / RAM
Downloads
Likes
LFM2.5 24B A2B MLX 4bit128K / 13.5 GB653524
LFM2.5 24B A2B MLX 5bit128K / 16.5 GB13276
LFM2.5 24B A2B MLX 6bit128K / 19.5 GB8282
LFM2.5 24B A2B OptiQ 4bit128K / 16.7 GB4521
...5 24B A2B Abliterated 4bit Mlx128K / 13.5 GB3761
...5 24B A2B Abliterated MLX 4bit128K / 13.5 GB3081
LFM2.5 24B A2B QAD MLX 4bit128K / 1.6 GB1911
LFM2 24B A2B MLX 8bit125K / 25.5 GB16518
LFM2 24B A2B MLX 4bit125K / 13.4 GB18310
LFM2 24B A2B MLX 6bit125K / 19.5 GB1048
Note: green Score (e.g. "73.2") means that the model is better than LiquidAI/LFM2.5-24B-A2B-MLX-8bit.