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.

Most downloadedNewestLargestLongest contextSmallest
IN
InternVL3_5 8B
Internvl · bartowski
Params8.19B
Context40K
Best fitQ8_0 10.4 GB
Vision
GE
gemma 4 E2B it
Gemma · AtomicChat
Params4.65B
Context128K
Best fitQ8_0 6.1 GB
ReasoningTool use
GE
Gemma 2 9b It
Gemma · bartowski
Params9.24B
Context8K
Best fitQ6_K_L 11.1 GB
GGUF
ON
OneJev 27B
Onejev · bartowski
Params27B
Context256K
Best fitIQ2_XS 11.5 GB
Vision
QW
Qwen3.8 Flash Next Coder cut240
Qwen · Schiltmans
Params56.6B
Context256K
Smallest4.5 GB
Coding
QW
Qwen3 14B
Qwen · ggml-org
Params14.8B
Context40K
Best fitQ4_K_M 10.8 GB
GGUF
CA
causallm_7b
TheBloke
Params7.72B
Context8K
Best fitQ6_K 11.1 GB
Tool useVision
TI
TIPO v2.1 1B A200M
Tipo · KBlueLeaf
Params1.01B
Context4K
Best fitF16 2.9 GB
CodingVision
SM
SmolLM2 135M
SmolLM · QuantFactory
Params135M
Context8K
Best fitQ2_K 0.8 GB
GGUF
MI
MiniCPM5 2B
MiniCPM · bartowski
Params2.52B
Context128K
Best fitBF16 5.9 GB
CodingTool use
QW
Qwen2.5 Coder 1.5B Instruct Abliterated
Qwen · bartowski
Params1.78B
Context32K
Best fitF16 4.3 GB
Coding
GL
GLM 5.3 Flash UNCENSORED mlx oQ8e DFlash2
GLM · Solstice-AI
Params1.17B
Context1024K
Best fitBF16 3.0 GB
Vision
MU
Muse Glimmer 30B
Muse · NANI-Nithin
Params28B
Context128K
Best fitQ2_K 11.6 GB
Tool useVision
MA
ms marco MiniLM L6 v2 Q8 0
Marco · sinjab
Params23M
Context512
Best fitQ8_0 0.6 GB
CodingEmbedding
GE
Gemma 3 12b It Heretic v2
Gemma · DreamFast
Params11.8B
Context128K
Best fitQ5_K_S 12.0 GB
CodingVision
GP
Gpt Oss 20b
gpt-oss · bartowski
Params20.9B
Context128K
Q4_K_M11.7 GB
GGUF
QW
Qwen3.5 9B Heretic v2
Qwen · filvyb
Params8.95B
Context256K
Best fitQ8_0 11.1 GB
Vision
GE
ThinkingCap gemma 4 26B A4B it
Gemma · bottlecapai
Params25.2B
Context256K
Q4_K_M17.0 GB
ReasoningTool useVision
NE
NeoHorse-1-9B
Neohorse · TokenRhythm
Params8.95B
Context256K
Best fitQ8_0 11.1 GB
ReasoningTool use
DE
deepseek-ai_deepseek-coder-6.7b-base
DeepSeek · TheBloke
Params6.74B
Context16K
Best fitQ8_0 12.0 GB
Coding
GR
Granite 4.1 3b
Granite · ibm-granite
Params3.4B
Context128K
Best fitBF16 8.0 GB
GGUF
GE
gemma-2-27b-it
Gemma · bartowski
Params27.2B
Context8K
Best fitIQ2_XS 12.0 GB
GGUF
JA
JackOD 9B Coder
Jackod · ManniX-ITA
Params8.95B
Context256K
Best fitQ8_0 11.1 GB
CodingReasoningTool use
DE
DeepSeek R1 Distill Qwen 14B Abliterated v2
DeepSeek · mradermacher
Params14.8B
Context128K
Best fitQ4_K_M 11.1 GB
Reasoning
MA
Maple Preview
Maple · deepgrove
Params20.2B
Context128K
Best fitTQ2_0 7.3 GB
CodingReasoning
DE
DeepSeek R1 Distill Qwen 14B Uncensored
DeepSeek · mradermacher
Params14.8B
Context128K
Best fitQ4_K_M 11.1 GB
Reasoning
GE
Huihui gemma 4 12B agentic fable5 abliterated
Gemma · huihui-ai
Params11.9B
Context256K
Best fitF16 2.4 GB
ReasoningTool use
GP
Gpt Oss 120b Uncensored
gpt-oss · mradermacher
Params117B
Context128K
Q4_K_M87.9 GB
Reasoning
QW
Qwen2.5 7B Instruct Uncensored
Qwen · mradermacher
Params7.62B
Context32K
Best fitQ8_0 9.1 GB
GGUF
QW
Qwen3.8 27B EfficientThink VAL APEX I MiniPlus V3
Qwen · IsValorum
Params27.3B
Context256K
Best fitQ4_K_M 1.8 GB
CodingReasoningVision

Other memory sizes

Keep browsing