IQ3_S GGUF models

A GGUF quantization format. Lower numbers mean smaller files and more quality loss. 45340 models have a IQ3_S file. Sizes run from 1M to 1030B parameters. The longest trained context is 102400K tokens.

Most downloadedNewestLargestLongest contextSmallest
BI
Bielik 7B Instruct v0.1
Bielik · mradermacher
Params7.24B
Context4K
Q4_K_M4.4 GB
GGUF
DO
DonutHole-8x7B
Donuthole · mradermacher
Params46.7B
Context32K
Q4_K_M28.4 GB
GGUF
NE
Neural AlphaMistral 7B
Neural · mradermacher
Params7.24B
Context32K
Q4_K_M4.4 GB
GGUF
QW
Qwen3 coder REAP 25B A3B Rust
Qwen · Em-80
Params25B
Context256K
Q4_K_M15.1 GB
Coding
LL
Hermes 2 Theta Llama 3 70B
Llama · mradermacher
Params70.6B
Context128K
Q4_K_M42.5 GB
Tool use
LL
DeepHermes 3 Llama 3 8B Preview
Llama · mradermacher
Params16.5B
Context128K
Q4_K_M10.0 GB
ReasoningTool use
ST
stabilityai japanese stablelm instruct gamma 7b
Stabilityai · RichardErkhov
Params7.24B
Context4K
Q4_K_M4.4 GB
GGUF
QW
Qwen3.5 40B RoughHouse Claude 4.6 Opus Polar Deckard Uncensored Heretic Thinking
Qwen · mradermacher
Params39.1B
Context256K
Q4_K_M23.9 GB
CodingReasoningRoleplay
MI
Mistral Small 24B Instruct 2501 Heretic
Mistral · mradermacher
Params23.6B
Context32K
Q4_K_M14.3 GB
GGUF
LI
Lil Beastie 4b
Lil · mradermacher
Params4.02B
Context256K
Q4_K_M2.5 GB
ReasoningTool use
WA
Wayfarer 2 12B
Wayfarer · LatitudeGames
Params12.2B
Context128K
Q4_K_M7.5 GB
Roleplay
NE
Neuron 4B Instruct
Neuron · mradermacher
Params4.02B
Context256K
Q4_K_M2.5 GB
Tool use
TO
tokyotech llm Swallow 7b instruct v0.1
Tokyotech · mmnga
Params6.83B
Context4K
Q4_K_M4.1 GB
GGUF
NE
Neuraljack 12B MoE
Neuraljack · mradermacher
Params12.9B
Context32K
Q4_K_M7.8 GB
GGUF
KI
RP King 12b
King · mradermacher
Params12.2B
Context1000K
Q4_K_M7.5 GB
Roleplay
BO
Borealis 270m
Borealis · mradermacher
Params268M
Context32K
Q4_K_M0.3 GB
GGUF
QW
Qwen2.5 14B Instruct Abliterated
Qwen · mradermacher
Params14.8B
Context32K
Q4_K_M9.0 GB
GGUF
LL
Colossal LLaMA 2 7b Base
Llama · mradermacher
Params7.04B
Context4K
Q4_K_M4.3 GB
GGUF
GR
Granite 4.1 8b Abliterated AND Disinhibited
Granite · mradermacher
Params8.38B
Context128K
Q4_K_M5.1 GB
GGUF
QW
QiMing Holos Plus Qwen3 8B
Qwen · mradermacher
Params8.19B
Context40K
Q4_K_M5.0 GB
GGUF
ST
stable-code-instruct-3b
Stable · lmstudio-community
Params2.8B
Context16K
Q4_K_M1.7 GB
Coding
LL
Mini Lama 8B Base
Llama · mradermacher
Params8.49B
Context256K
Q4_K_M5.2 GB
Vision
SP
Spark X2.6 4B Base
Spark · mradermacher
Params4.11B
Context1024K
Q4_K_M2.6 GB
ReasoningTool use
LL
llama 65b instruct
Llama · mradermacher
Params65B
Context2K
Q4_K_M39.3 GB
GGUF
MI
dolphin-2.9.1-mixtral-1x22b
Mistral · InferenceIllusionist
Params22.2B
Context64K
Q4_K_M13.3 GB
CodingTool use
SE
SearchQwen2.5 7B
Searchqwen · mradermacher
Params7.62B
Context32K
Q4_K_M4.7 GB
Tool use
PL
Pluto i1
Pluto · mradermacher
Params8.95B
Context256K
Q4_K_M5.6 GB
CodingReasoningTool use
GI
Gille StrangeMerges 50 7B slerp
Gille · RichardErkhov
Params7.24B
Context32K
Q4_K_M4.4 GB
GGUF
YI
dolphin-2.9.1-yi-1.5-34b
Yi · mradermacher
Params34.4B
Context8K
Q4_K_M20.7 GB
CodingTool use
MI
General Stories Mistral 7B
Mistral · mradermacher
Params7.24B
Context32K
Q4_K_M4.4 GB
GGUF

Other quants

Keep browsing