Meta, abliterated by mlabonne
Llama 3.1 8B Abliterated
Llama 3.1 8B Abliterated is Meta's Llama 3.1 8B as abliterated by mlabonne, whose 2024 write-up made the method popular.
Rank#8 / 8
#8Editors' rankingSize#8 / 8
8BTotal parametersPrice
FreeOpen weights you run yourselfRuns on
8 GB GPU or laptopAbout 5 GB at 4-bitOur pick for a laptop or an 8 GB GPU.
Input
Model id
mlabonne/Meta-Llama-3.1-8B-Instruct-abliteratedIt is far behind the 2026 models on reasoning.
Run it yourself
SizeAbout 5 GB at 4-bit
HardwareAn 8 GB GPU or a laptop
LicenseLlama 3.1 Community License
Load a GGUF build in Ollama, LM Studio or llama.cpp. When a job outgrows one GPU, run DeepSeek V4.1 Flash Abliterated on Silk Compute.