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.
GE
Models Google Gemma 3 4b It
Gemma · MaziyarPanahi
Params3.88B
Context128K
Best fitQ4_K_S 4.0 GB
QW
Qwen3.5 27B
Qwen · bartowski
Params27.3B
Context256K
Q4_K_M18.0 GB
NE
NVIDIA-Nemotron-3.5-Super-VL-120B-A12B
Nemotron · ggml-org
Params1.66B
Context1024K
Best fitQ4_0 2.9 GB
QW
Qwen3.5-122B-A10B
Qwen · unsloth
Params122B
Context256K
Q4_K_M76.5 GB
GE
Gemma-3-1B-It
Gemma · unsloth
Params1B
Context32K
Best fitBF16 2.7 GB
ST
stable diffusion v1 5
Stable · second-state
Params1.07B
Context—
Best fitF16 2.9 GB
AN
anthracite org magnum 12b v2
Anthracite · RichardErkhov
Params12.2B
Context—
Best fitIQ3_XS 3.1 GB
MI
models--mistralai--Mistral-7B-Instruct-v0.3
Mistral · MaziyarPanahi
Params7.25B
Context32K
Best fitIQ2_XS 3.8 GB
JI
Jina Embeddings v5 Text Nano Retrieval
Jina · jinaai
Params212M
Context8K
Best fitF16 1.2 GB
MI
Ministral 3 3B Instruct 2512
Mistral · mistralai
Params3.43B
Context256K
Best fitQ5_K_M 3.8 GB
QW
Qwen3.6 27B
Qwen · bartowski
Params27.3B
Context256K
Q4_K_M18.0 GB
QW
Qwen3.5 2B
Qwen · bartowski
Params1.94B
Context256K
Best fitQ8_0 3.0 GB
JI
Jina Embeddings v5 Text Small Retrieval
Jina · jinaai
Params596M
Context40K
Best fitF16 2.6 GB
VO
Voxtral Mini 3B 2507
Voxtral · ggml-org
Params4B
Context128K
Best fitQ4_K_M 4.0 GB
QW
Qwen3 1.7B
Qwen · bartowski
Params2.03B
Context32K
Best fitQ8_0 3.6 GB
LF
LFM2.5 2.6B DSpark Draft v1
Lfm · LiquidAI
Params328M
Context125K
Best fitF16 1.2 GB
CY
Cydonia 24B v4.3
Cydonia · bartowski
Params23.6B
Context128K
Q4_K_M14.3 GB
QW
Qwen2.5 Coder 7B Instruct Abliterated
Qwen · bartowski
Params7.62B
Context32K
Best fitQ2_K 4.0 GB
QW
Qwen3 0.6B
Qwen · ggml-org
Params752M
Context40K
Best fitF16 2.9 GB
MI
MiniCPM5 1B
MiniCPM · openbmb
Params1.08B
Context128K
Best fitF16 2.9 GB
QW
Qwen2.5 VL 7B Instruct
Qwen · mradermacher
Params7.62B
Context125K
Best fitQ2_K 4.0 GB
GR
ibm granite granite 20b code base
Granite · RichardErkhov
Params1.9B
Context—
Best fitIQ3_M 3.2 GB
LL
CreativeWriter Llama3.2 3B
Llama · theprint
Params3.21B
Context128K
Best fitQ5_K_M 3.8 GB
LL
models--MaziyarPanahi--Llama-3-8B-Instruct-32k-v0.1
Llama · MaziyarPanahi
Params8B
Context8K
Best fitIQ1_M 3.7 GB
GE
Gemma 4 E4B It
Gemma · bartowski
Params7.5B
Context128K
Q4_K_M5.4 GB
GE
Huihui gemma 4 12B it qat q4 0 unquantized abliterated
Gemma · huihui-ai
Params11.9B
Context256K
Best fitBF16 2.4 GB
LL
Models Meta Llama Llama 3.2 1B Instruct
Llama · MaziyarPanahi
Params1.24B
Context128K
Q4_K_M0.8 GB
FR
FrogNano 4B 2609
Frognano · bartowski
Params4.33B
Context256K
Best fitIQ3_M 4.0 GB
QW
Models Qwen Qwen3 4B Instruct 2507
Qwen · MaziyarPanahi
Params4B
Context256K
Best fitQ3_K_L 3.9 GB
LL
Llama 3.2 3B Instruct
Llama · bartowski
Params3.61B
Context128K
Best fitQ4_K_L 4.0 GB