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.
YI
models--01-ai--Yi-1.5-6B-Chat
Yi · MaziyarPanahi
Params6.06B
Context4K
Best fitQ6_K 6.0 GB
GE
Models Google Gemma 3 12b It
Gemma · MaziyarPanahi
Params11.8B
Context128K
Q4_K_M7.3 GB
LL
models--MaziyarPanahi--Llama-3-8B-Instruct-64k
Llama · MaziyarPanahi
Params8B
Context63K
Q4_K_M4.9 GB
MI
Models Mistralai Mistral Nemo Instruct 2407
Mistral · MaziyarPanahi
Params12.2B
Context1000K
Q4_K_M7.5 GB
LL
Models Meta Llama Llama 3.3 70B Instruct
Llama · MaziyarPanahi
Params70.6B
Context128K
Best fitQ4_K_M 45.7 GB
MI
Mistral-7B-Instruct-v0.3
Mistral · thesven
Params7B
Context32K
Best fitQ8_0 9.3 GB
MI
models--mistralai--mathstral-7B-v0.1
Mistral · MaziyarPanahi
Params7.25B
Context32K
Q4_K_M4.4 GB
TP
tpn004 cand v3f aim
Tpn · tatjr13
Params23.6B
Context128K
Smallest14.7 GB
TP
tpn004 cand v4f l1im
Tpn · tatjr13
Params23.6B
Context128K
Smallest18.3 GB
MI
Rocinante X 12B v1
Mistral · TheDrummer
Params12.2B
Context128K
Best fitQ8_0 14.9 GB
TP
tpn004 cand v3f l1im
Tpn · tatjr13
Params23.6B
Context128K
Smallest18.3 GB
QW
Qwen3.8 27B TWIN TURBO Fable Cold Fusion 709 ULTRA HERETIC Uncensored NM DAU NEO MTP
Qwen · DavidAU
Params27B
Context256K
Best fitQ8_0 32.9 GB
QW
Qwen3 VL 32B Instruct ultra uncensored heretic H3
Qwen · nif0
Params25.2B
Context256K
Best fitQ4_K_M 17.1 GB
LL
Llama 3.2 1B Instruct
Llama · hugging-quants
Params1.24B
Context128K
Best fitQ4_K_M 1.6 GB
LL
Models Meta Llama Meta Llama 3.1 70B Instruct
Llama · MaziyarPanahi
Params70.6B
Context128K
Best fitQ4_K_M 45.7 GB
TP
tpn004 cand v4f aim
Tpn · tatjr13
Params23.6B
Context128K
Smallest14.7 GB
TP
tpn004 cand v1f aim
Tpn · tatjr13
Params23.6B
Context128K
Smallest14.7 GB
QW
Qwen Image 2.1 PE T2I Heretic
Qwen · pottokao
Params8.95B
Context256K
Best fitQ8_0 11.1 GB
LL
Llama-4-Scout-17B-16E-Instruct
Llama · unsloth
Params108B
Context10240K
Best fitIQ3_XXS 47.8 GB
OP
Tinyopenjev For Testing
Openjev · ggml-org
Params33M
Context256K
Best fitQ8_0 0.6 GB
MI
MiniMax H3 encoder
Minimax · joeygambino
Params32.8B
Context256K
Best fitQ5_K_M 25.9 GB
QW
Qwen3.5 35B A3B
Qwen · bartowski
Params35.5B
Context256K
Best fitQ8_0 39.0 GB
OR
Ornith 1.5 9B
Ornith · AtomicChat
Params8.95B
Context256K
Best fitBF16 19.5 GB
GL
Glm-4.6V-Flash
GLM · unsloth
Params9.4B
Context128K
Best fitBF16 19.6 GB
QW
models--Qwen--Qwen2-7B-Instruct
Qwen · MaziyarPanahi
Params7.62B
Context32K
Q4_K_M4.7 GB
OR
Orion 26B A4B v1.1
Orion · bartowski
Params25.2B
Context256K
Best fitQ8_0 35.4 GB
TR
Translategemma 4b It
Translategemma · mradermacher
Params3.88B
Context128K
Best fitF16 9.4 GB
QW
Qwen3 ASR 1.7B
Qwen · ggml-org
Params2.03B
Context64K
Best fitBF16 5.5 GB
QW
Qwen2.5 Coder 1.5B Instruct
Qwen · bartowski
Params1.54B
Context32K
Best fitF16 3.8 GB
MI
Models Mistralai Mistral Small Instruct 2409
Mistral · MaziyarPanahi
Params22.2B
Context32K
Q4_K_M13.3 GB