TheDrummer/Fallen-Command-A-111B-v1

111B • Updated Mar 24, 2025 • 7 • 26

Note Recommended IQ2_XXS for offloading or Q5_K_M for high VRAM. The darkest model in existence.

SicariusSicariiStuff/Assistant_Pepe_8B

266k • Updated 24 days ago • 206 • 68

Note Creative and uncensored. 8B version outperforms 70B. Not good at following system prompts. For RP you can take fallen command output, wrap it in < think > tags, and feed it to pepe. Useful for dataset generation. 32B doesn't work on my PC for some reason.

DarkArtsForge/Morax-24B-v2

Text Generation • Updated 1 day ago • 81 • 5

Note Favorite 24B, extremely dark model

OccultAI/Qliphoth-12B-v1

Text Generation • 12B • Updated 2 days ago • 110 • • 1

Note Favorite RP model currently

mlabonne/gemma-3-27b-it-abliterated

Image-Text-to-Text • 27B • Updated Mar 21, 2025 • 86.9k • • 324

Note Better than G4 at following complex system prompts

concedo/Huihui-GLM-4.5-Air-abliterated-lossytensors

Text Generation • 110B • Updated Oct 13, 2025 • 10 • 5

Note This model works well alongside Gemma models. Some refusals with complex agentic system prompts.

DarkArtsForge/huihui-ai_Huihui-GLM-4.5-Air-abliterated-Q3_K_M-GGUF

110B • Updated 5 days ago • 152

Note best version of the above model for 64GB RAM

EldritchLabs/MN-12B-Mag-Mell-R1-Uncensored-Scale1.2

Text Generation • 12B • Updated Mar 5 • 9 • • 5

Note My favorite general purpose 12B model. Also good at RP. Smarter in some ways than 24B.

Naphula/Delirium-v1-abliterated

Text Generation • 9B • Updated Nov 19, 2025 • 3 • 1

Note Very creative, unhinged writer once you add the recommended deslop tokenbans

Naphula/Goetia-8B-v1

Text Generation • 8B • Updated Apr 1 • 19 • • 7

Note Tied for favorite 8B

OccultAI/Adversary-8B-v1a

8B • Updated 11 days ago • 46 • 2

Note Tied for favorite 8B

OccultAI/Red-Pilled-Morpheus-7B-v1

Text Generation • 7B • Updated Feb 22 • 25 • 2

Note Pairs well with Assistant Pepe 8B

Naphula/G4-Runic-Oarfish-26B-A4B-v1.2

Image-Text-to-Text • 25B • Updated 19 days ago • 491 • 11

Note Not as good at writing as Tiger Gemma although it is smarter overall. G4 has a lot of slop in general.

TheDrummer/Tiger-Gemma-9B-v3

9B • Updated Oct 4, 2024 • 19 • 61

Note IQ3_S fits entirely on 8GB VRAM. Tiger 9B works surprisingly well for AGENTIC tasks. If you NEED SPEED fully uncensored, this is the best budget model, except maybe 26B A4B for higher context

Naphula/DarkForest-20B-v2.0-Erebus-Edition

20B • Updated Nov 23, 2025 • 83 • 11

Note Creative and uncensored RP, dark and unhinged, the OG of this list

EldritchLabs/Human-Like-Mistral-Nemo-Instruct-2407-MPOA

Text Generation • 12B • Updated Mar 5 • 6 • • 2

Note Useful for adding emojis to dataset generations.