GGUF models that fit in 4 GB
79024 GGUF language models have at least one quant whose estimated memory use is 4 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 1280000K tokens.
UM
Umievo-itr012-Gleipnir-7B
Umievo · mmnga
Params7.24B
Context8K
Best fitIQ2_S 3.9 GB
KI
Kisoku 1.6b Chat
Kisoku · mradermacher
Params1.6B
Context128K
Best fitQ6_K 2.2 GB
HU
Hunyuan 1.8B Instruct
Hunyuan · bartowski
Params1.79B
Context256K
Best fitQ8_0 2.9 GB
GE
Gemma 4 E4B It Apostate
Gemma · mradermacher
Params8.13B
Context128K
Q4_K_M5.0 GB
AU
Audar TTS V1 Turbo
Audar · audarai
Params1.64B
Context32K
Best fitQ8_0 2.5 GB
QI
QiMing Holos Plus 8B
Qiming · mradermacher
Params8B
Context128K
Best fitIQ2_XXS 4.0 GB
NO
North Mini Code 1.0 Uncensored Heretic
North · mradermacher
Params30.5B
Context488K
Q4_K_M18.6 GB
LL
samantha-dolphin-llama-3
Llama · mradermacher
Params7.24B
Context32K
Best fitIQ2_S 3.9 GB
HE
Hemmingway 1 Heretic MTP V3 Final
Hemmingway · mradermacher
Params27.3B
Context256K
Q4_K_M16.8 GB
QW
Qwen3.8 9B
Qwen · Atomic-Germ
Params9.2B
Context256K
Q4_K_M5.8 GB
GR
Granite 4.0 350m
Granite · ibm-granite
Params352M
Context32K
Best fitBF16 2.1 GB
CY
Cydonia 24B v4.1
Cydonia · bartowski
Params23.6B
Context128K
Q4_K_M14.3 GB
MI
Ministral 3 3B Reasoning 2512
Mistral · bartowski
Params3.43B
Context256K
Best fitQ5_K_L 3.9 GB
OR
Ornith 1.0 9B MTP
Ornith · protoLabsAI
Params9.2B
Context256K
Best fitQ8_0 4.0 GB
KO
Kohlibri 40B
Kohlibri · mradermacher
Params40.3B
Context256K
Q4_K_M24.5 GB
MI
dolphin 2.2.1 mistral 7b i1
Mistral · mradermacher
Params7.24B
Context32K
Best fitIQ2_S 3.9 GB
SK
Skyfall 31B v4.2 Abliterix
Skyfall · mradermacher
Params31.4B
Context128K
Q4_K_M19.0 GB
MI
Dolphin Mistral GLM 4.7 Flash 24B Venice Edition Thinking Uncensored
Mistral · mradermacher
Params23.6B
Context32K
Q4_K_M14.3 GB
LL
Hermes 3 Llama 3.2 3B
Llama · mradermacher
Params3.21B
Context128K
Best fitQ5_K_M 3.8 GB
ME
Meltemi 7B v1.5
Meltemi · mradermacher
Params7.48B
Context32K
Best fitIQ2_S 4.0 GB
QW
Qwen3 Coder 30B A3B Instruct Heretic
Qwen · mradermacher
Params30.5B
Context256K
Q4_K_M18.6 GB
RO
RoLlama3-8b-Instruct
Rollama · legraphista
Params8B
Context8K
Best fitIQ2_XXS 4.0 GB
CO
ConfigurableHermes 7B i1
Configurablehermes · mradermacher
Params7.24B
Context32K
Best fitIQ2_S 3.9 GB
ST
StartLux Decision v1 0.8B Q8 0
Startlux · startlux-models
Params752M
Context256K
Best fitQ8_0 1.7 GB
QW
Qwen3.5 2B Claude 4.6 OS Auto Variable HERETIC UNCENSORED THINKING
Qwen · mradermacher
Params1.88B
Context256K
Best fitQ8_0 2.9 GB
GE
Gemma 2 2b It
Gemma · lmstudio-community
Params2.61B
Context8K
Best fitQ6_K 3.5 GB
ME
Mellum2 12B A2.5B Instruct
Mellum · bartowski
Params12.1B
Context128K
Q4_K_M8.2 GB
MI
Mistral Medium 3.5 128B Eschaton Uncensored
Mistral · mradermacher
Params125B
Context256K
Q4_K_M74.9 GB
QW
Qwen3 30B A3B Gemini Pro High Reasoning 2507 ABLITERATED UNCENSORED
Qwen · mradermacher
Params30.5B
Context256K
Q4_K_M18.6 GB
ST
stable-code-instruct-3b
Stable · mradermacher
Params2.8B
Context16K
Best fitIQ2_XXS 4.0 GB