LLM EXPLORER 61,652 MODELS INDEXED

LFM2.5 8B A1B Abliterated by DuoNeural

By DuoNeural · 191 downloads
HF Repo ↗ Copy URL

LFM2.5 8B A1B Abliterated is an open-source language model by DuoNeural. Features: 8b LLM, VRAM: 17GB, Context: 125K, License: apache-2.0, LLM Explorer Score: 0.25.

  Abliteration   Conversational   Duoneural   En   Hybrid-attention   Lfm2 moe   Liquid-ai   Moe   Reasoning   Region:us   Safetensors   Sharded   Tensorflow   Uncensored
Model Card on HF 🤗: https://huggingface.co/DuoNeural/LFM2.5-8B-A1B-Abliterated 

LFM2.5 8B A1B Abliterated Parameters and Internals

LLM NameLFM2.5 8B A1B Abliterated
Repository 🤗https://huggingface.co/DuoNeural/LFM2.5-8B-A1B-Abliterated 
Base Model(s)  liquid-ai/lfm-2.5-8b-a1b   liquid-ai/lfm-2.5-8b-a1b
Model Size8b
Required VRAM17 GB
Updated2026-09-19
MaintainerDuoNeural
Model Typelfm2_moe
Model Files  4.7 GB: 1-of-4   4.9 GB: 2-of-4   4.7 GB: 3-of-4   2.7 GB: 4-of-4
Supported Languagesen
Model ArchitectureLfm2MoeForCausalLM
Licenseapache-2.0
Context Length128000
Model Max Length128000
Transformers Version5.9.0
Tokenizer ClassTokenizersBackend
Padding Token<|pad|>
Vocabulary Size128000

Best Alternatives to LFM2.5 8B A1B Abliterated

Best Alternatives
Context / RAM
Downloads
Likes
LFM2.5 8B A1B125K / 16.9 GB84755670
LFM2.5 8B A1B UltraData L2 49M125K / 16.9 GB2931
...inal ToolBench Full SFT 1Epoch125K / 16.9 GB88672
LFM2.5 8B A1B Base125K / 16.9 GB420129
LFM2 8B A1B125K / 16.7 GB30570370
LFM2.5 8B A1B Opus Distil125K / 16.9 GB21275
LFM2.5 8B A1B125K / 16.9 GB124013
LFM2.5 8B A1B NVFP4125K / 6.9 GB4607
LFM2.5 8B A1B MLX Bf16125K / 16.8 GB3598
Supertron2.1 8B A1B125K / 17 GB966
Note: green Score (e.g. "73.2") means that the model is better than DuoNeural/LFM2.5-8B-A1B-Abliterated.