2026/11/18The November 2026 edition ranks 6 APIs and 8 models
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' ranking
Size#8 / 8
8BTotal parameters
Price
FreeOpen weights you run yourself
Runs on
8 GB GPU or laptopAbout 5 GB at 4-bit

Our pick for a laptop or an 8 GB GPU.

Input
Model idmlabonne/Meta-Llama-3.1-8B-Instruct-abliterated

It is far behind the 2026 models on reasoning.

Run it yourself

Weightsmlabonne/Meta-Llama-3.1-8B-Instruct-abliterated
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.

Ranked for

Other abliterated models