GGUF models that fit in 48 GB

201364 GGUF language models have at least one quant whose estimated memory use is 48 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
QW
Models Qwen Qwen3 30B A3B
Qwen · MaziyarPanahi
Params30.5B
Context40K
Best fitQ6_K 26.4 GB
GGUF
GE
Gemma 3 4b It
Gemma · lmstudio-community
Params3.88B
Context128K
Best fitQ8_0 5.8 GB
Vision
QW
Qwen3 Omni 30B A3B Instruct
Qwen · ggml-org
Params30.5B
Context64K
Best fitQ8_0 33.8 GB
Vision
QW
Models Qwen Qwen3 32B
Qwen · MaziyarPanahi
Params32.8B
Context40K
Best fitQ6_K 29.5 GB
GGUF
QW
Huihui Qwen3.6 27B Abliterated
Qwen · huihui-ai
Params27.3B
Context256K
Best fitQ8_0 31.7 GB
Vision
ST
stories15M MOE
Stories · ggml-org
Params8M
Context256
Smallest0.0 GB
GGUF
LL
Llama 3.3 70B Instruct
Llama · bartowski
Params70.6B
Context128K
Best fitQ4_K_L 46.5 GB
GGUF
AR
Artemis 31B v1.2
Artemis · bartowski
Params30.7B
Context256K
Best fitQ3_K_M 47.4 GB
Vision
QW
Qwen3 Embedding 4B
Qwen · Qwen
Params4B
Context40K
Best fitF16 9.8 GB
Embedding
NE
NVIDIA Nemotron 3.5 Bolt 74B A9B
Nemotron · unsloth
Params74B
Context256K
Best fitQ3_K_XL 39.6 GB
Vision
FA
Jiqing tiny random falcon
Falcon · RichardErkhov
Params—
Context—
Best fitIQ3_M 0.5 GB
GGUF
QW
Qwen3.5 0.8B
Qwen · ggml-org
Params773M
Context256K
Best fitBF16 2.5 GB
Vision
HU
Cyber Tiel Coder 35B A3B
Huihui · peculiar-ragdoll
Params34.7B
Context256K
Best fitQ8_K_XL 39.6 GB
CodingReasoningTool use
FR
FrontTerrain-HARP-1B
Frontterrain · frontterrain
Params123M
Context512
Best fitF16 0.8 GB
Coding
QW
Qwen3-VL-32B Instruct MiniMax-H3 L0-49 Q4_K_M
Qwen · nif0
Params25.2B
Context256K
Best fitQ4_K_M 17.1 GB
Vision
LL
Llama 3.3 70B Instruct Abliterated
Llama · bartowski
Params70.6B
Context128K
Best fitQ4_1 47.5 GB
GGUF
GE
Winnow 12B
Gemma · EldanRing
Params423M
Context256K
Best fitBF16 25.4 GB
Vision
KO
Kolibri 1
Kolibri · Hob-forge
Params78B
Context256K
Best fitQ3_K_M 38.9 GB
ReasoningTool use
QW
Qwen_Qwen3 VL 8B Instruct
Qwen · lmstudio-community
Params8.19B
Context256K
Best fitQ8_0 10.4 GB
Vision
IN
Index Translate 9B
Index · IndexTeam
Params9.2B
Context256K
Best fitF16 20.0 GB
Vision
SM
SmolLM2 135M Instruct
SmolLM · bartowski
Params135M
Context8K
Best fitF16 1.0 GB
GGUF
QW
Huihui Qwythos 9B Claude Mythos 5 1M Abliterated MTP
Qwythos · huihui-ai
Params9.2B
Context1024K
Best fitBF16 20.0 GB
ReasoningTool useVision
WA
Wald-4B v1.2
Wald · org2ai
Params4.2B
Context256K
Best fitQ8_0 6.1 GB
Vision
GE
Gemma-3N-E2B-It
Gemma · unsloth
Params4.46B
Context32K
Best fitF16 9.9 GB
Vision
QW
Qwen3-32B
Qwen · unsloth
Params32.8B
Context40K
Best fitQ8_K_XL 42.1 GB
GGUF
QW
Qwen3.8 27B pi
Qwen · bytkim
Params27B
Context256K
Best fitQ8_0 29.3 GB
CodingReasoningTool use
GE
Gemma 2 2b It
Gemma · bartowski
Params2.61B
Context8K
Best fitF32 11.8 GB
GGUF
GR
ibm granite granite 20b code instruct
Granite · RichardErkhov
Params1.9B
Context—
Best fitIQ3_XS 10.2 GB
Coding
LL
Llama 3.2 3B Instruct
Llama · lmstudio-community
Params3.21B
Context128K
Best fitQ8_0 4.9 GB
GGUF
LL
Llama3.3 8B Instruct Thinking Heretic Uncensored Claude 4.5 Opus High Reasoning
Llama · mradermacher
Params8B
Context128K
Best fitQ6_K 8.2 GB
ReasoningRoleplay

Other memory sizes

Keep browsing