GGUF models that fit in 12 GB
171517 GGUF language models have at least one quant whose estimated memory use is 12 GB or less at 8K context. Each card shows the largest quant that fits. Sizes run from 1M to 20260731B parameters. The longest trained context is 4194304K tokens.
QW
Qwen3 4B Abliterated
Qwen · bartowski
Params4B
Context40K
Best fitBF16 9.8 GB
SK
Skyfall 31B v4.2
Skyfall · bartowski
Params31.4B
Context128K
Best fitIQ2_XXS 12.0 GB
QW
Qwopus3 5 9B Coder Flash Test2 MTP HF
Qwen · Jackrong
Params9.2B
Context256K
Best fitQ8_0 11.4 GB
NS
NSFW flash
Nsfw · UnfilteredAI
Params2.24B
Context4K
Best fitQ4_K_M 4.0 GB
QW
Qwen3.8 35B A3B Distill
Qwen · IsValorum
Params34.7B
Context256K
Best fitQ8_0 3.2 GB
ST
StartLux-Decision-v1.1-4B-AutoThink
Startlux · startlux-models
Params4.2B
Context256K
Best fitQ8_0 6.1 GB
MI
MiniCPM5 2B heretic abliterated
MiniCPM · Abiray
Params2.52B
Context128K
Best fitQ8_0 3.5 GB
EM
Embeddinggemma 300m Qat Q4_0 Unquantized
Embeddinggemma · unsloth
Params303M
Context2K
Best fitF32 1.9 GB
QW
Qwen3.8-27B
Qwen · kingjones777
Params26.9B
Context256K
Best fitQ8_0 5.8 GB
QW
Qwen2.5 3B Instruct Uncensored
Qwen · mradermacher
Params3.09B
Context32K
Best fitF16 7.0 GB
MI
Ministral-3-3B-Reasoning-2512
Mistral · unsloth
Params3.43B
Context256K
Best fitBF16 8.2 GB
MI
Ministral 8B Instruct 2410 TEST
Mistral · bartowski
Params8B
Context32K
Best fitQ8_0 10.2 GB
NU
NuExtract3
Nuextract · numind
Params4.33B
Context256K
Best fitQ8_0 6.2 GB
PA
PaddleOCR VL 1.6
Paddleocr · LunarOilRig
Params467M
Context128K
Best fitQ4_K_M 1.0 GB
HY
Hy-Mt2-7B
unsloth
Params7.5B
Context256K
Best fitQ8_K_XL 11.2 GB
QW
Qwen3.6 35B A3B MTP APEX I MiniPlus V2.1 Abliterated
Qwen · IsValorum
Params34.7B
Context256K
Best fitQ8_0 3.2 GB
GE
Gemmasutra Mini 2B v1
Gemma · TheDrummer
Params2.61B
Context8K
Best fitF32 11.8 GB
QW
Qwen3.5-9B
Qwen · jc-builds
Params8.95B
Context256K
Best fitQ4_K_M 7.3 GB
QW
Qwen3-Coder-30B-A3B-Instruct-1M
Qwen · unsloth
Params30.5B
Context1024K
Best fitIQ2_XXS 11.7 GB
GE
google gemma 3 12b it qat q4 0 gguf small fix
Gemma · Dampfinchen
Params11.8B
Context128K
Best fitQ4_0 10.6 GB
QW
Huihui Qwen3.5 9B Abliterated
Qwen · mradermacher
Params8.95B
Context256K
Best fitQ6_K 8.9 GB
OR
Ornith 1.0 35B
Ornith · unsloth
Params34.7B
Context256K
Best fitIQ1_S 11.7 GB
QW
Qwen3 8B
Qwen · Aldaris
Params8.19B
Context40K
Best fitQ4_K_M 6.7 GB
QW
Huihui Qwen3.8 27B Coder390 EfficientThink Abliterated
Qwen · mradermacher
Params27.3B
Context256K
Q4_K_M16.8 GB
QW
Qwen2.5 14B
Qwen · bartowski
Params14.8B
Context128K
Best fitQ4_K_L 11.7 GB
QW
Qwen3.6 12B IQ Ultra Heretic Uncensored Thinking V2 Hightop
Qwen · KevinJK51
Params11.7B
Context256K
Best fitQ6_K 10.9 GB
PH
phi 4
Phi · microsoft
Params14.7B
Context16K
Best fitQ4_1 11.4 GB
LA
latxa-7b-v1.2
Latxa · mradermacher
Params6.74B
Context4K
Best fitQ6_K 10.3 GB
GE
gemma 4 E4B it heretic
Gemma · Abiray
Params7.5B
Context128K
Best fitQ8_0 9.9 GB
AP
APUS Reflex 9B
Apus · apus-ailab
Params8.95B
Context256K
Best fitQ8_0 11.1 GB