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.

Most downloadedNewestLargestLongest contextSmallest
QW
Qwen2.5 VL 3B Instruct Abliterated
Qwen · huihui-ai
Params3.09B
Context125K
Best fitQ4_K_M 2.7 GB
Vision
QW
Qwen3 0.6B Heretic Heresy Edition
Qwen · mradermacher
Params596M
Context40K
Best fitF16 2.6 GB
ReasoningTool use
XY
XYZ Aquila Mini
Xyz · bartowski
Params34.7B
Context256K
Q4_K_M21.4 GB
Vision
LF
LFM3 1.7B
Lfm · LiquidAI
Params1.73B
Context256K
Best fitQ8_0 4.0 GB
ReasoningTool use
AR
BAAI AREX 2
Arex · bartowski
Params27B
Context256K
Q4_K_M17.2 GB
ReasoningTool useVision
BA
BarryFutureman NeuralTurdusVariant1 7B
Barryfutureman · RichardErkhov
Params7B
Context—
Best fitQ3_K_S 4.0 GB
GGUF
QW
Qwen3.6 35B A3B Uncensored Heretic
Qwen · mradermacher
Params34.7B
Context256K
Q4_K_M21.2 GB
GGUF
MI
MiniCPM V 4_6
MiniCPM · openbmb
Params752M
Context256K
Best fitF16 2.4 GB
Vision
GE
Gemma-3-4B-It-Qat
Gemma · unsloth
Params3.88B
Context128K
Best fitIQ4_NL 4.0 GB
Vision
QW
Qwen3.5-0.8B
Qwen · weareapexcreators
Params773M
Context256K
Best fitQ4_K_M 1.5 GB
Vision
QW
Qwen2 VL 2B Instruct
Qwen · bartowski
Params1.54B
Context32K
Best fitF16 3.8 GB
Vision
FL
Flint-40B-A2.5B-abliterated
Flint · dima6312
Params39.8B
Context256K
Q4_K_M24.2 GB
CodingReasoningTool use
QW
Huihui Qwen3.5 35B A3B Abliterated
Qwen · mradermacher
Params34.7B
Context256K
Q4_K_M21.2 GB
GGUF
QW
Therapy 3.8
Qwen · Verdugie
Params26.9B
Context256K
Q4_K_M16.5 GB
ReasoningTool use
LF
LFM2.5-350M
Lfm · weareapexcreators
Params354M
Context125K
Best fitQ8_0 1.4 GB
GGUF
QW
Qwen3.8-27B-DFlash2
Qwen · perplexity-ai
Params27.3B
Context256K
Best fitQ8_0 3.8 GB
Vision
FL
RecursiveSI flint
Flint · bartowski
Params39.8B
Context256K
Q4_K_M25.3 GB
GGUF
LF
LFM2 VL 1.6B
Lfm · LiquidAI
Params1.17B
Context125K
Best fitF16 3.9 GB
Vision
LL
VAGOsolutions Llama 3 SauerkrautLM 70b Instruct
Llama · RichardErkhov
Params6B
Context—
Best fitQ3_K_L 4.0 GB
GGUF
PA
PaddleOCR VL 1.5
Paddleocr · PaddlePaddle
Params467M
Context128K
Smallest0.9 GB
Vision
GE
gemma-2b-it
Gemma · lmstudio-ai
Params2.51B
Context8K
Best fitQ8_0 3.3 GB
GGUF
LF
LFM2.5 230M LocalAI
Lfm · weareapexcreators
Params230M
Context125K
Best fitQ8_0 1.2 GB
GGUF
JI
jina reranker v1 tiny en
Jina · mradermacher
Params33M
Context8K
Best fitQ2_K 0.6 GB
CodingEmbedding
MI
Mistral Small 4 119B 2603
Mistral · bartowski
Params119B
Context1024K
Q4_K_M72.6 GB
Vision
QW
Qwen3.5 27B Claude 4.6 Opus Reasoning Distilled
Qwen · mradermacher
Params27B
Context256K
Q4_K_M16.5 GB
Reasoning
SP
Spark X2.6 4B
Spark · bartowski
Params4.11B
Context1024K
Best fitIQ4_XS 4.0 GB
GGUF
QW
Qwen3.5-2B
Qwen · weareapexcreators
Params1.94B
Context256K
Best fitQ5_K_M 2.4 GB
Vision
LL
Llama 3.2 3B Instruct abliterated
Llama · QuantFactory
Params3.2B
Context128K
Best fitQ5_1 3.9 GB
GGUF
QW
Qwen3.5 2B Claude 4.6 Opus Reasoning Distilled
Qwen · Jackrong
Params1.88B
Context256K
Best fitQ8_0 2.9 GB
ReasoningVision
MI
MiniCPM5 2B LocalAI
MiniCPM · weareapexcreators
Params2.52B
Context128K
Best fitQ8_0 3.5 GB
GGUF

Other memory sizes

Keep browsing